7HM_mouse_human_comparison.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Mouse / human comparison\n",
    "\n",
    "For this notebook **you need to run the 4M and 4H notbeooks previously!!**.\n",
    "\n",
    "For part of the analysis **you need to run the 0A notbeook** too. This part is not strictly necessary to be run, but instead it helps me do part of some gene analyses quicker. If you have trouble loading content from this notebook, ignore it because the analysis will be the same.\n",
    "\n",
    "In this notebook we are going to analyse the similarities and differences between mouse and human skin fibroblast populations. We already know that mouse and human skin are different between them, but we want to know how much of this translates into the transcriptomic realm of single cell.\n",
    "\n",
    "To do this analysis we are going to do two analyses. \n",
    "* Choose datasets with human and mouse samples from the same lab (e.g., Boothby, Vorstandlechner) and try to find overlaps in the populations.\n",
    "* Map human/mouse genes between them using an homology databe (MGI) and try to find similarities between the list of markers of human and mouse populations. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import scanpy as sc\n",
    "import scanpy.external as sce\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import triku as tk\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl\n",
    "from tqdm.notebook import tqdm\n",
    "import scipy.sparse as spr\n",
    "import matplotlib.cm as cm\n",
    "import networkx as nx\n",
    "from matplotlib import pylab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "!pip install munkres\n",
    "from munkres import Munkres"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# local imports and imports from other notebooks\n",
    "from cellassign import assign_cats\n",
    "from fb_functions import make_gene_scoring_with_expr, plot_score_graph, plot_UMAPS_gene, plot_adata_cluster_properties\n",
    "%store -r dict_colors_human\n",
    "%store -r dict_colors_mouse\n",
    "\n",
    "dict_colors_human_mouse = {**dict_colors_human , **dict_colors_mouse}\n",
    "\n",
    "%store -r seed\n",
    "%store -r magma\n",
    "%store -r data_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "%store -r dict_make_gene_scoring_robust\n",
    "%store -r dict_make_gene_scoring_axis_robust"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# FROM NOTEBOOK 0A!!\n",
    "%store -r df_human_genes_codes\n",
    "%store -r df_mouse_genes_codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "%store -r plot_params\n",
    "\n",
    "pylab.rcParams.update(plot_params)\n",
    "pd.set_option('display.max_columns', None)\n",
    "pd.options.display.float_format = \"{:,.2f}\".format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def join_fbs_adatas(adata_full, adata_fb):\n",
    "    cell_types = adata_full.obs['assigned_cats'].copy().astype(str)\n",
    "    intersect_idx = np.intersect1d(adata_fb.obs_names, adata_full.obs_names)\n",
    "    cell_types[intersect_idx] = [f'fibro_{i}' for i in adata_fb[intersect_idx].obs['cluster']]\n",
    "    adata_full.obs['full_cell_type'] = cell_types.astype('category')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-05-02T14:04:59.807789Z",
     "iopub.status.busy": "2022-05-02T14:04:59.807311Z",
     "iopub.status.idle": "2022-05-02T14:04:59.838952Z",
     "shell.execute_reply": "2022-05-02T14:04:59.838234Z",
     "shell.execute_reply.started": "2022-05-02T14:04:59.807735Z"
    },
    "tags": []
   },
   "source": [
    "## Creating the mouse-human gene homology dictionary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# !cd results && wget http://www.informatics.jax.org/downloads/reports/HOM_AllOrganism.rpt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('results/HOM_AllOrganism.rpt', sep='\\t')\n",
    "df = df[df['Common Organism Name'].isin(['mouse, laboratory', 'human'])][['DB Class Key', 'Common Organism Name', 'Symbol']].reset_index(drop=True)\n",
    "\n",
    "list_DB = set(df['DB Class Key'].values)\n",
    "\n",
    "dict_mouse_human = {}\n",
    "for el in tqdm(list_DB):\n",
    "    df_sub = df[df['DB Class Key'] == el].sort_values(by='Common Organism Name')\n",
    "    if len(df_sub) == 2:\n",
    "        dict_mouse_human[df_sub.iloc[1, 2]] = df_sub.iloc[0, 2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "dict_human_mouse = dict(zip(dict_mouse_human.values(), dict_mouse_human.keys()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparison of UMAPs of populations\n",
    "\n",
    "In this section we are going to compare mouse and human datasets of \n",
    "* Datasets from the same laboratory\n",
    "* Datasets from diferent laboratories (more confirmatory)\n",
    "\n",
    "To do this we are going to translate the mouse into human genes and get the subset of genes that have homology and are present in the human adata."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Boothby"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "boothby_2021_dir = data_dir + '/boothby_2021'\n",
    "boothby_2021_ctrl_mouse_fb = sc.read(boothby_2021_dir + '/boothby_2021_ctrl_mouse_fb_robust.h5')\n",
    "boothby_2021_ctrl_human_fb = sc.read(boothby_2021_dir + '/boothby_2021_ctrl_human_fb_robust.h5')\n",
    "\n",
    "boothby_2021_ctrl_mouse_fb_raw = sc.read(boothby_2021_dir + '/boothby_2021_mouse_ctrl_mouse.h5')\n",
    "boothby_2021_ctrl_human_fb_raw = sc.read(boothby_2021_dir + '/boothby_2021_ctrl_human.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "boothby_2021_ctrl_mouse_fb.X = boothby_2021_ctrl_mouse_fb_raw[boothby_2021_ctrl_mouse_fb.obs_names, boothby_2021_ctrl_mouse_fb.var_names].X.copy()\n",
    "boothby_2021_ctrl_human_fb.X = boothby_2021_ctrl_human_fb_raw[boothby_2021_ctrl_human_fb.obs_names, boothby_2021_ctrl_human_fb.var_names].X.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "adata_mouse_genes, adata_human_genes = boothby_2021_ctrl_mouse_fb.var_names, boothby_2021_ctrl_human_fb.var_names\n",
    "mouse_selected_genes, human_homolog, human_mouse_gene = [], [], []\n",
    "\n",
    "for i in adata_mouse_genes:\n",
    "    if i in dict_mouse_human:\n",
    "        if (dict_mouse_human[i] in adata_human_genes) & (dict_mouse_human[i] not in human_homolog):\n",
    "            mouse_selected_genes.append(i); human_homolog.append(dict_mouse_human[i]); human_mouse_gene.append(f'{dict_mouse_human[i]} | {i}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "boothby_2021_ctrl_human_fb, boothby_2021_ctrl_mouse_fb = boothby_2021_ctrl_human_fb[:, human_homolog], boothby_2021_ctrl_mouse_fb[:, mouse_selected_genes]\n",
    "boothby_2021_ctrl_human_fb.var_names, boothby_2021_ctrl_mouse_fb.var_names = human_mouse_gene, human_mouse_gene\n",
    "boothby_2021_ctrl_human_mouse_fb = sc.AnnData.concatenate(boothby_2021_ctrl_human_fb, boothby_2021_ctrl_mouse_fb, batch_categories=['human', 'mouse'], batch_key='organism')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.log1p(boothby_2021_ctrl_human_mouse_fb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.pca(boothby_2021_ctrl_human_mouse_fb, random_state=seed, n_comps=50)\n",
    "sce.pp.harmony_integrate(boothby_2021_ctrl_human_mouse_fb, key='Internal sample identifier', max_iter_harmony=50)\n",
    "sc.pp.neighbors(boothby_2021_ctrl_human_mouse_fb, use_rep='X_pca_harmony', n_neighbors=int(0.5 * len(boothby_2021_ctrl_human_mouse_fb) ** 0.5 // 4), metric='cosine')\n",
    "tk.tl.triku(boothby_2021_ctrl_human_mouse_fb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "boothby_2021_ctrl_human_mouse_fb.var_names[boothby_2021_ctrl_human_mouse_fb.var['highly_variable'] == True]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.umap(boothby_2021_ctrl_human_mouse_fb, min_dist=0.2, random_state=seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "boothby_2021_ctrl_human_mouse_fb.obs['cluster'] = boothby_2021_ctrl_human_mouse_fb.obs['cluster'].astype('category')\n",
    "boothby_2021_ctrl_human_mouse_fb.uns['cluster_colors'] = [dict_colors_human_mouse[i] if i in dict_colors_human_mouse else '#bcbcbc' for  i in boothby_2021_ctrl_human_mouse_fb.obs['cluster'].cat.categories]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.subsample(boothby_2021_ctrl_human_mouse_fb, fraction=1, random_state=0, copy=False)\n",
    "sc.pl.umap(boothby_2021_ctrl_human_mouse_fb, color=['cluster', 'organism'], legend_loc='on data', frameon=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.rank_genes_groups(boothby_2021_ctrl_human_mouse_fb, groupby='organism')\n",
    "sc.pl.umap(boothby_2021_ctrl_human_mouse_fb, color=['organism'], legend_loc='on data')\n",
    "sc.pl.rank_genes_groups_tracksplot(boothby_2021_ctrl_human_mouse_fb, groupby='organism', n_genes=10, dendrogram=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "adata_regress = sc.pp.regress_out(boothby_2021_ctrl_human_mouse_fb, 'organism', n_jobs=32, copy=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.pca(adata_regress, random_state=seed, n_comps=50)\n",
    "sce.pp.harmony_integrate(adata_regress, key='Internal sample identifier', max_iter_harmony=50)\n",
    "sc.pp.neighbors(adata_regress, use_rep='X_pca_harmony', n_neighbors=int(0.5 * len(adata_regress) ** 0.5 // 4), metric='cosine')\n",
    "\n",
    "sc.tl.umap(adata_regress)\n",
    "sc.pl.umap(adata_regress, color=['cluster', 'organism'], legend_loc='on data', frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vorstandlechner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "vorstandlechner_2021_dir = data_dir + '/Vorstandlechner_2021'\n",
    "vorstandlechner_2021_ctrl_human_fb = sc.read(f\"{vorstandlechner_2021_dir}/vors_2021_ctrl_human_fb_robust.h5\")\n",
    "vorstandlechner_2021_ctrl_mouse_fb = sc.read(f\"{vorstandlechner_2021_dir}/vorstandlechner_2021_ctrl_mouse_fb_robust.h5\")\n",
    "\n",
    "vorstandlechner_2021_ctrl_human_fb_raw = sc.read(f\"{vorstandlechner_2021_dir}/vorstandlechner_2021_ctrl_human.h5\")\n",
    "vorstandlechner_2021_ctrl_mouse_fb_raw = sc.read(f\"{vorstandlechner_2021_dir}/vorstandlechner_2021_ctrl_mouse.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "vorstandlechner_2021_ctrl_human_fb.X = vorstandlechner_2021_ctrl_human_fb_raw[vorstandlechner_2021_ctrl_human_fb.obs_names, vorstandlechner_2021_ctrl_human_fb.var_names].X.copy()\n",
    "vorstandlechner_2021_ctrl_mouse_fb.X = vorstandlechner_2021_ctrl_mouse_fb_raw[vorstandlechner_2021_ctrl_mouse_fb.obs_names, vorstandlechner_2021_ctrl_mouse_fb.var_names].X.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "adata_mouse_genes, adata_human_genes = vorstandlechner_2021_ctrl_mouse_fb.var_names, vorstandlechner_2021_ctrl_human_fb.var_names\n",
    "mouse_selected_genes, human_homolog, human_mouse_gene = [], [], []\n",
    "\n",
    "for i in adata_mouse_genes:\n",
    "    if i in dict_mouse_human:\n",
    "        if (dict_mouse_human[i] in adata_human_genes) & (dict_mouse_human[i] not in human_homolog):\n",
    "            mouse_selected_genes.append(i); human_homolog.append(dict_mouse_human[i]); human_mouse_gene.append(f'{dict_mouse_human[i]} | {i}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "vorstandlechner_2021_ctrl_human_fb, vorstandlechner_2021_ctrl_mouse_fb = vorstandlechner_2021_ctrl_human_fb[:, human_homolog], vorstandlechner_2021_ctrl_mouse_fb[:, mouse_selected_genes]\n",
    "vorstandlechner_2021_ctrl_human_fb.var_names, vorstandlechner_2021_ctrl_mouse_fb.var_names = human_mouse_gene, human_mouse_gene\n",
    "vorstandlechner_2021_ctrl_human_mouse_fb = sc.AnnData.concatenate(vorstandlechner_2021_ctrl_human_fb, vorstandlechner_2021_ctrl_mouse_fb, batch_categories=['human', 'mouse'], batch_key='organism')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.log1p(vorstandlechner_2021_ctrl_human_mouse_fb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.pca(vorstandlechner_2021_ctrl_human_mouse_fb, random_state=seed, n_comps=50)\n",
    "sce.pp.harmony_integrate(vorstandlechner_2021_ctrl_human_mouse_fb, key='Internal sample identifier', max_iter_harmony=50)\n",
    "sc.pp.neighbors(vorstandlechner_2021_ctrl_human_mouse_fb, use_rep='X_pca_harmony', n_neighbors=int(0.5 * len(vorstandlechner_2021_ctrl_human_mouse_fb) ** 0.5 // 4), metric='cosine')\n",
    "tk.tl.triku(vorstandlechner_2021_ctrl_human_mouse_fb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "vorstandlechner_2021_ctrl_human_mouse_fb.var_names[vorstandlechner_2021_ctrl_human_mouse_fb.var['highly_variable'] == True]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, min_dist=0.2, random_state=seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "vorstandlechner_2021_ctrl_human_mouse_fb.obs['cluster'] = vorstandlechner_2021_ctrl_human_mouse_fb.obs['cluster'].astype('category')\n",
    "vorstandlechner_2021_ctrl_human_mouse_fb.uns['cluster_colors'] = [dict_colors_human_mouse[i] if i in dict_colors_human_mouse else '#bcbcbc' for  i in vorstandlechner_2021_ctrl_human_mouse_fb.obs['cluster'].cat.categories]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.subsample(vorstandlechner_2021_ctrl_human_mouse_fb, fraction=1, random_state=0, copy=False)\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['cluster', 'organism'], legend_loc='on data', frameon=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.rank_genes_groups(vorstandlechner_2021_ctrl_human_mouse_fb, groupby='organism')\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['organism'], legend_loc='on data')\n",
    "sc.pl.rank_genes_groups_tracksplot(vorstandlechner_2021_ctrl_human_mouse_fb, groupby='organism', n_genes=10, dendrogram=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "adata_regress = sc.pp.regress_out(vorstandlechner_2021_ctrl_human_mouse_fb, 'organism', n_jobs=32, copy=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.pca(adata_regress, random_state=seed, n_comps=50)\n",
    "sce.pp.harmony_integrate(adata_regress, key='Internal sample identifier', max_iter_harmony=50)\n",
    "sc.pp.neighbors(adata_regress, use_rep='X_pca_harmony', n_neighbors=int(0.5 * len(adata_regress) ** 0.5 // 4), metric='cosine')\n",
    "\n",
    "sc.tl.umap(adata_regress)\n",
    "sc.pl.umap(adata_regress, color=['cluster', 'organism'], legend_loc='on data', frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Showing DEGs of human VS mouse populations in Vorstandlechner et al 2021"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.subsample(vorstandlechner_2021_ctrl_human_mouse_fb, fraction=1, random_state=0, copy=False)\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['cluster'], legend_loc='on data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb[vorstandlechner_2021_ctrl_human_mouse_fb.obs['organism']=='mouse'], color=['cluster'], legend_loc='on data')\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb[vorstandlechner_2021_ctrl_human_mouse_fb.obs['organism']=='human'], color=['cluster'], legend_loc='on data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.leiden(vorstandlechner_2021_ctrl_human_mouse_fb, resolution=0.01)\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['leiden'], legend_loc='on data')\n",
    "sc.tl.rank_genes_groups(vorstandlechner_2021_ctrl_human_mouse_fb, groupby='leiden')\n",
    "sc.pl.rank_genes_groups_tracksplot(vorstandlechner_2021_ctrl_human_mouse_fb, n_genes=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pp.subsample(vorstandlechner_2021_ctrl_human_mouse_fb, fraction=1, random_state=0, copy=False)\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['cluster'], legend_loc='on data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb[vorstandlechner_2021_ctrl_human_mouse_fb.obs['organism']=='mouse'], color=['cluster'], legend_loc='on data')\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb[vorstandlechner_2021_ctrl_human_mouse_fb.obs['organism']=='human'], color=['cluster'], legend_loc='on data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sc.tl.leiden(vorstandlechner_2021_ctrl_human_mouse_fb, resolution=0.01)\n",
    "sc.pl.umap(vorstandlechner_2021_ctrl_human_mouse_fb, color=['leiden'], legend_loc='on data')\n",
    "sc.tl.rank_genes_groups(vorstandlechner_2021_ctrl_human_mouse_fb, groupby='leiden')\n",
    "sc.pl.rank_genes_groups_tracksplot(vorstandlechner_2021_ctrl_human_mouse_fb, n_genes=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusions\n",
    "Datasets do not match. When reading the methods they point out that they use a restricted set of genes to create then do the UMAP. I don't feel confortable with that strategy, so I prefer to do a more \"unbiased\" analysis based on the markers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparison of gene patterns between populations\n",
    "In this section we are going to take the gene markers from human and mouse, and do human-mouse, mouse-mouse and human-human comparisons between the set of markers. From this we expect to find similarities betwen both parties and establish a homology model of the skin fibroblasts.\n",
    "\n",
    "To do this analysis we are going to:\n",
    "* Get the top N markers for each population in mouse and human (also in human-human and mouse-mouse) and compute the Jaccard index between them. \n",
    "    * To find the \"best\" N we are going to compute for an array of N the matrix the jaccard matrix and get the trace with highest values using the hungarian algorithm. The sum of the trace is stored. A higher sum would imply a general higher overlap between markers and, therefore, a more relevant analysis. From the array of values, we are going to get the best \"rounded\" N.\n",
    "* With the top N we are going to get the Jaccard matrix and plot it with a heatmap and a clustergram.\n",
    "* From there, we are going to select the best matches and analyze the relationship between human and mouse populations\n",
    "\n",
    "We do the human-human and mouse-mouse comparisons to get a \"control\" view. We expect more gene overlap between clusters of the same axis than of different ones. Also, we get a sense of the expected overlaps, to see how this is reflected in mouse.\n",
    "\n",
    "To make the joining as \"easy\" as possible, we are going to use the subset of genes that have a mouse-human homology only. This might limit the extent of the analysis but, in general, we find a fair overlap, and this will make the dataset mapping and batch effect correction much easier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "%store -r dict_make_gene_scoring_cluster_robust_human\n",
    "%store -r dict_make_gene_scoring_cluster_robust_mouse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def get_df_overlap(dict_1, dict_2, N=100, translate=True):\n",
    "    df_overlap = pd.DataFrame(0, index=dict_1.keys(), columns=dict_2.keys())\n",
    "    \n",
    "    for cluster_name_1, cluster_df_1 in dict_1.items():\n",
    "        for cluster_name_2, cluster_df_2 in dict_2.items():\n",
    "            gene_list_1 = set(cluster_df_1.index[:N])\n",
    "            gene_list_2_unchanged = cluster_df_2.index[:N]\n",
    "            if translate:\n",
    "                gene_list_2 = set([dict_mouse_human[i] if i in dict_mouse_human else i for i in gene_list_2_unchanged])\n",
    "            else:\n",
    "                gene_list_2 = set(gene_list_2_unchanged)\n",
    "\n",
    "            overlap = len(gene_list_1 & gene_list_2) / len(gene_list_1 | gene_list_2)\n",
    "            if overlap == 1:\n",
    "                overlap = 0\n",
    "                \n",
    "            df_overlap.loc[cluster_name_1, cluster_name_2] = overlap\n",
    "    \n",
    "    return df_overlap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_best_N(dict_1, dict_2, N_min=10, N_max=300, translate=True):\n",
    "    list_N, list_total = [],[] \n",
    "    for N in tqdm(range(N_min, N_max)):\n",
    "        df_jaccard_1_2 = get_df_overlap(dict_1, dict_2, N=N, translate=translate)\n",
    "\n",
    "        m = Munkres()\n",
    "        indexes = m.compute(1-df_jaccard_1_2.values)\n",
    "        total = 0\n",
    "        for row, column in indexes:\n",
    "            value = df_jaccard_1_2.values[row][column]\n",
    "            total += value\n",
    "\n",
    "        list_N.append(N)\n",
    "        list_total.append(total)\n",
    "\n",
    "    plt.plot(list_N, list_total)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_heatmap(df_jaccard_1_2, dict_colors_1, dict_colors_2, figsize=(12,8), ticklabelsize=15, diag=False, annot=True, cmap='magma'):\n",
    "    fig, ax = plt.subplots(1, 1, figsize=figsize)\n",
    "    if diag:\n",
    "        mask = np.eye(df_jaccard_1_2.shape[0], dtype=bool)\n",
    "        sns.heatmap(df_jaccard_1_2, annot=annot, fmt='.2f', ax=ax, mask=mask, cmap=cmap)\n",
    "        ax.set_facecolor(\"#989898\")\n",
    "    else:\n",
    "        sns.heatmap(df_jaccard_1_2, annot=annot, fmt='.2f', ax=ax, cmap=cmap)\n",
    "    [t.set_color(dict_colors_1[t.get_text()]) for t in ax.xaxis.get_ticklabels()]; [t.set_color(dict_colors_2[t.get_text()]) for t in ax.yaxis.get_ticklabels()]\n",
    "    ax.set_xticklabels(ax.get_xticklabels(),  weight='bold', size=ticklabelsize); ax.set_yticklabels(ax.get_yticklabels(),  weight='bold', size=ticklabelsize, va='center')\n",
    "    None # To avoid plotting stuff on screen\n",
    "    \n",
    "def plot_clustermap(df_jaccard_1_2, dict_colors_1, dict_colors_2, figsize=(12,8), ticklabelsize=15, diag=False):\n",
    "    if diag:\n",
    "        mask = np.eye(df_jaccard_1_2.shape[0], dtype=bool)\n",
    "        cg = sns.clustermap(df_jaccard_1_2, mask=mask)\n",
    "        ax = cg.ax_heatmap\n",
    "        ax.set_facecolor(\"#989898\")\n",
    "    else:\n",
    "        cg = sns.clustermap(df_jaccard_1_2, )\n",
    "        ax = cg.ax_heatmap\n",
    "        \n",
    "    [t.set_color(dict_colors_1[t.get_text()]) for t in ax.xaxis.get_ticklabels()]; [t.set_color(dict_colors_2[t.get_text()]) for t in ax.yaxis.get_ticklabels()]\n",
    "    ax.set_xticklabels(ax.get_xticklabels(),  weight='bold', size=ticklabelsize); ax.set_yticklabels(ax.get_yticklabels(),  weight='bold', size=ticklabelsize, va='center')\n",
    "    None # To avoid plotting stuff on screen"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def print_common_genes(dict_1, dict_2, cluster_1, cluster_2, translate=True, N=150):\n",
    "    gene_list_1 = set(dict_1[cluster_1].index[:N])\n",
    "    gene_list_2_unchanged = dict_2[cluster_2].index[:N]\n",
    "    if translate:\n",
    "        gene_list_2 = set([dict_mouse_human[i] if i in dict_mouse_human else i for i in gene_list_2_unchanged])\n",
    "    else:\n",
    "        gene_list_2 = set(gene_list_2_unchanged)\n",
    "        \n",
    "    return sorted(list(gene_list_1 & gene_list_2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-05-03T09:52:46.091273Z",
     "iopub.status.busy": "2022-05-03T09:52:46.090756Z",
     "iopub.status.idle": "2022-05-03T09:52:46.125286Z",
     "shell.execute_reply": "2022-05-03T09:52:46.124363Z",
     "shell.execute_reply.started": "2022-05-03T09:52:46.091211Z"
    },
    "tags": []
   },
   "source": [
    "### Human-human comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "plot_best_N(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_human, N_min=10, N_max=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_jaccard_human_human = get_df_overlap(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_human, N=150)\n",
    "plot_heatmap(df_jaccard_human_human, dict_colors_human, dict_colors_human, figsize=(12,7), ticklabelsize=15, diag=True, cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# sorted\n",
    "df_jaccard_human_human_sorted = df_jaccard_human_human.loc[['A2', 'A3', 'A1', 'A4', 'B1', 'B3', 'B2', 'B4', 'C2', 'C3', 'C1', 'C5', 'D1', 'D2', 'E1'], \n",
    "                                                           ['A2', 'A3', 'A1', 'A4', 'B1', 'B3', 'B2', 'B4', 'C2', 'C3', 'C1', 'C5', 'D1', 'D2', 'E1']]\n",
    "\n",
    "df_jaccard_human_human_sorted[df_jaccard_human_human_sorted < 0.03] = np.nan\n",
    "\n",
    "plot_heatmap(df_jaccard_human_human_sorted, dict_colors_human, dict_colors_human, figsize=(12,7), ticklabelsize=15, annot=True, cmap='Blues')\n",
    "\n",
    "plot_heatmap(df_jaccard_human_human_sorted, dict_colors_human, dict_colors_human, figsize=(12,7), ticklabelsize=15, annot=False, cmap='Blues')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comments on human-human\n",
    "* It is clear that A3 is a bridge between A1-A4 and A2. A2 and A1 are clearly different clusters, and A3 is more \"favorable\" to A1 than to A2.\n",
    "    * A1 - A4 (no A3): **ACKR3**, ADGRD1, C1QTNF3, **CA12**, CD151, CD248, **CD34**, **CD55**, **CD70**, CHRDL1, **CLEC3B**, CTHRC1, CYB5R3, **DBN1**, DPP4, **EMILIN2**, **FAM180B**, **FAP**, FBLN2, **FBN1**, FSTL1, **GALNT15**, **GPX3**, **IGFBP6**, **ISLR**, **LGR5**, LIMS2, MATN4, MFAP5, MMP2, PALM, PAMR1, PCOLCE2, PIGT, PLAC9, PPIC, **PRG4**, **SCARA5**, SEMA3C, SLC29A1, SLITRK4, **SLPI**, SSC5D, TGFBR3, TIMP2, TNXB, VASN, VAT1 [48 total, **19**]\n",
    "    * A1 - A3 (no A4): ADAMTSL1, AEBP1, ANGPTL1, ANGPTL5, **AOX1**, **ARFGEF3**, C1QTNF3, CD248, CERCAM, CLEC3B, **COL14A1**, COL1A1, COL1A2, **CORIN**, CPE, CPVL, CREB5, CRYAB, CTHRC1, CTSB, CTSK, CYB5R3, CYBRD1, DCN, DPP4, ECM1, ELN, FBLN1, FBLN2, FBN1, FKBP9, **GLRB**, HEXA, ISLR, ITIH5, LOX, LRFN5, LRRC2, MMP2, MMP27, NUCB2, OLFML3, **OMD**, PCOLCE2, PDGFRL, PI16, PLAC9, PLD3, PODN, PPIC, QPCT, RECK, SEMA3B, SERPINF1, SFRP2, SMOC2, SPARC, SSC5D, **SVEP1**, THBS2, THBS3, TNXB, WISP2, XG [64 total, **7**]\n",
    "    * A4 - A1 - A3 (all 3): **C1QTNF3**, **CD248**, CLEC3B, CTHRC1, CYB5R3, **DPP4,** **FBLN2**, FBN1, ISLR, MMP2, **PCOLCE2**, PLAC9, PPIC, SSC5D, **TNXB** [15 total, **6**]\n",
    "    * A2 - A3: CD9, COL5A1, COL6A2, HSPB7, **ISM1**, **LEPR**, NBL1, NCKAP5, **RSPO1**, SFRP2, TWIST2, **WIF1**, WNT16 [13 total, **X**]\n",
    "\n",
    "\n",
    "* A2 has a \"shared\" transcriptomic profile with C and E axis! It is also apparent form UMAPs and connectivity graphs. However, its relatively low similarity profile with C1, C3 and E1 makes it a somewhat \"independent\" population.\n",
    "    * A2-C1: APELA, CCDC3, **COL21A1**, COL27A1, COL3A1, **COL7A1**, **COL6A3**, CPXM2, **DKK3**, DUSP4, **LAMC3**, ROBO2, ROR2, **SEMA5A**, **SPON1**, **STMN1**, **THSD4**, **TMEM119**  [18 total, **10**]\n",
    "    * A2-C3: **ADGRE2**, AK5, ANTXR1, AQP1, C4orf48, COL3A1, **COL5A1**, **COL6A1**, **COL6A3**, **COL7A1**, **COMP**, CPXM2, DKK3, F13A1, LBH, **LOXL2**, MFAP2, **PTK7**, RAB31, RGS3, SCARF2, SEMA5A, **SPON1**, SUGCT, TANC2, TCF4, TMEM119, TNC, UNC5B  [29 total, **9**]\n",
    "    * A2-E1: ACVR2A, AK5, **AQP1**, CDC42EP3, **CMKLR1**, COL21A1, DKK3, **DUSP6**, HS3ST6, ID1, **IGFBP2**, LAMC3, **MAP2**, PLCB1, PREX1, **RGCC**, SAMD5, SEMA5A, SPON1, **SPRY1**, TGFBI, UNC5B  [22 total, **7**]\n",
    "\n",
    "\n",
    "* B axis seems quite independent from the rest of axes, although it is related to T1 and, also, B1 with D1. \n",
    "    * It is also claear that B3 acts as a bridge between B1, B2 and B4. The bridge is more clear between B3 and B2 than B3 and B1. B1 and B2 are not related, and B4 is slightly related to B2, but not B1 (although they appear more related in UMAPs and graphs).\n",
    "        * B3 - B1 (not B2): ADAMTS4, APOE, **BIRC3**, **CCL2**, **CXCL1**, **CXCL2**, **CXCL3**, **GEM**, **HAS2**, ICAM1, IRF1, IRF8, NFKB2, NFKBIA, PHLDA1, PPP1R15A, RELB, SOCS3, **TNFAIP3**, UGCG, ZNF267 [21 total, **8**]\n",
    "        * B3 - B2 (not B1): ABCC3, ACHE, ADAMDEC1, **ADRA2A**, **ANKRD29**, APLNR, APOE, B2M, C1RL-AS1, **C3**, **CCL19**, **CD74**, CDX1, CH25H, **CLSTN3**, COX4I2, **CRB2**, **CTSC**, CTSH, **CX3CL1**, CXCL12, CYP7B1, ENC1, EXOC3L4, FLT3LG, FXYD6, GGT5, GUCY1A1, GUCY1B1, HES4, HLA-B, HLA-DRB1, HLA-F, **ICAM2**, **IGFBP3**, IGFBP7, IL15, IL33, **IL34**, IRF8, ITGA8, **JAK3**, LIFR, MICALL2, MSC, **NLGN4X**, NPB, ODF3B, OLFM2, OSMR, P2RY14, PGM5, PGM5-AS1, PKP2, PLA2G4C, PLEKHH2, POPDC2, PSMB9, **PTGDS**, PTPRT, RARRES2, RASSF4, **RBP5**, SCN4B, SDK1, SLC51A, SLC9A3R2, **SLCO2B1**, **ST8SIA1**, TMEM132C, **TMEM150C**, TMEM176A, TMEM176B, TNFSF10, **TNFSF13B**, **TYMP**, UBD, VEGFA, WIPF3 [79 total, **21**]\n",
    "        * B1 - B3 - B2 (all 3): **APOE**, **IRF8** [2 total, **2**]\n",
    "    * B1, similar to A2, shows a *more* independent transcriptomic profile with its neighbour B clusters.\n",
    "    * B4 is more similar to B2 than to the rest of B clusters\n",
    "        * B4 - B2: **ABCA10**, **ABCA8**, ADH1B, APOC1, APOE, C1orf54, C3, C7, COL4A4, CXCL12, CYGB, DEPTOR, EBF1, EFEMP1, EPHX1, **FXYD6**, GGT5, HEYL, IFITM1, IGFBP7, IL33, NRP1, PDE5A, PRRX1, PTCH2, RARRES2, SERPINF2, SLIT2, TMEM176A, VWA1 [30 total, **3**]\n",
    "    * B4 shares a decent set of genes with D1, D2 and E1\n",
    "        * B4 - D1: A2M, **ABCA8**, **ABCA9**, ABLIM1, AKR1C1, **APOD**, BNC2, CNTN4, CRISPLD1, **CYGB**, EGFR, EPS8, FLRT2, FMO1, FMO2, FMO3, FOXS1, **GPC3**, **IGFBP7**, INMT, **ITM2A**, LGALS3BP, **LSP1**, NGFR, NR2F2, OAF, PEAR1, PLPPR4, SFRP4, SHISA3, SOX5, SPTBN1, TFPI, **VIT**  [34 total, **9**]\n",
    "        * B4 - D2: A2M, **ABCA8**, ADAMTSL3, BNC2, CCL13, **CHRDL1**, CNTN4, COL4A2, DACT1, EGFR, FHL2, FLRT2, FOXS1, **IGFBP6**, IGFBP7, INMT, ITGA7, **ITM2A**, ITM2B, NGFR, NOTCH3, NR2F2, OAF, PDZRN4, PEAR1, PHLDA3, **RARRES2**, SFRP4, SMIM10, SPTBN1, TFPI, TGFBR2, **TXNIP**, **VIT**  [35 total, **7**]\n",
    "        * B4 - E1: A2M, **APOD**, ARHGDIB, CALD1, **COL15A1**, CRLF1, DACT1, EPHA3, FHL2, FMO2, **GSN**, **IGF1**, INMT, **ITM2A**, ITM2B, **LSP1**, **MGP**, OAF, PEAR1, **PLA2G5**, PLBD1, PRSS23, RARRES2, RELN, SLC16A4, SPON2, **SPRY1**, **STMN2**, SULT1A1, **TIMP3**, TSHZ2, TXNIP, VWA1  [33 total, **11**]\n",
    "    * B1 also shares some similarities with D1:\n",
    "        * B1 - D1: AMD1, **CHD1**, CYCS, DNAJA1, DUSP5, EGR3, ERF, KLF5, LDLR, **MAFF**, NFE2L2, PHLDA1 [12 total, **2**]\n",
    "\n",
    "\n",
    "* The C, D and E axes, although they have a substructure, they are more correlated between them than A or B axes.\n",
    "    * C1 and C3 are quite related, and C3 specially is the bridge cluster among the C axis (with C1 specially, then with C2 and then with C5). \n",
    "    * D1 and D2 are also very related.\n",
    "    * E1 axis interacts to some extent with the all C and D axes, but also with A2 and B4; which makes it a great candidate to analyse where and how it is located within the dermis. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-05-03T09:52:46.091273Z",
     "iopub.status.busy": "2022-05-03T09:52:46.090756Z",
     "iopub.status.idle": "2022-05-03T09:52:46.125286Z",
     "shell.execute_reply": "2022-05-03T09:52:46.124363Z",
     "shell.execute_reply.started": "2022-05-03T09:52:46.091211Z"
    },
    "tags": []
   },
   "source": [
    "### Mouse-mouse comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "plot_best_N(dict_make_gene_scoring_cluster_robust_mouse, dict_make_gene_scoring_cluster_robust_mouse, N_min=10, N_max=300, translate=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_jaccard_mouse_mouse = get_df_overlap(dict_make_gene_scoring_cluster_robust_mouse, dict_make_gene_scoring_cluster_robust_mouse, N=150, translate=False)\n",
    "plot_heatmap(df_jaccard_mouse_mouse, dict_colors_mouse, dict_colors_mouse, figsize=(12,7), ticklabelsize=15, diag=True, cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# sorted\n",
    "df_jaccard_mouse_mouse_sorted = df_jaccard_mouse_mouse.loc[['z2', 'z1', 'x2', 'x1', 'x/y', 'y3', 'y1', 'y2', 'y4', 'y5', 'v1', 'w/x', 'w4', 'w3', 'w1', 'w2', 'w5'], \n",
    "                                                           ['z2', 'z1', 'x2', 'x1', 'x/y', 'y3', 'y1', 'y2', 'y4', 'y5', 'v1', 'w/x', 'w4', 'w3', 'w1', 'w2', 'w5']]\n",
    "\n",
    "df_jaccard_mouse_mouse_sorted[df_jaccard_mouse_mouse_sorted < 0.05] = np.nan\n",
    "\n",
    "plot_heatmap(df_jaccard_mouse_mouse_sorted, dict_colors_mouse, dict_colors_mouse, figsize=(12,7), ticklabelsize=15, annot=False, cmap='Blues')\n",
    "plot_heatmap(df_jaccard_mouse_mouse_sorted, dict_colors_mouse, dict_colors_mouse, figsize=(12,7), ticklabelsize=15, annot=True, cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "cluster_1, cluster_2, cluster_3 = 'w3', 'w/x', 'x2'\n",
    "\n",
    "list_genes_int_1 = print_common_genes(dict_make_gene_scoring_cluster_robust_mouse, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2, translate=False, N=150)\n",
    "\n",
    "list_genes_int_2 = print_common_genes(dict_make_gene_scoring_cluster_robust_mouse, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_2, cluster_3, translate=False, N=150)\n",
    "\n",
    "list_genes_int = [i for i in list_genes_int_1 if i in list_genes_int_2]\n",
    "\n",
    "print(', '.join(list_genes_int), f'[{len(list_genes_int)} total, *X*]')\n",
    "\n",
    "aaa = df_mouse_genes_codes.loc[[i for i in list_genes_int if i in df_mouse_genes_codes.index]]\n",
    "aaa.loc[np.array([cluster_1 in i for i in aaa['code'].values]) & np.array([cluster_2 in i for i in aaa['code'].values]) & np.array([cluster_3 in i for i in aaa['code'].values])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "cluster_1, cluster_2 = 'x2', 'w/x'\n",
    "# cluster_not = 'y5'\n",
    "\n",
    "list_genes_int = print_common_genes(dict_make_gene_scoring_cluster_robust_mouse, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2, translate=False, N=150)\n",
    "print(', '.join(list_genes_int), f'[{len(list_genes_int)} total, *X*]')\n",
    "\n",
    "aaa = df_mouse_genes_codes.loc[[i for i in list_genes_int if i in df_mouse_genes_codes.index]]\n",
    "aaa.loc[np.array([cluster_1 in i for i in aaa['code'].values]) & np.array([cluster_2 in i for i in aaa['code'].values])  ]\n",
    "#         & np.array([cluster_not not in i for i in aaa['code'].values])]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comments on mouse-mouse\n",
    "* z1 - z2: 1700019D03Rik, *Aif1l*, *Akr1c18*, *Aldh1a3*, Anxa1, Anxa3, Axl, Basp1, C3, Car8, Cd248, Cd55, Chrdl1, Cmah, Col14a1, Creb5, Dbn1, Ddr2, *Dpp4*, Ebf2, Efemp1, *Efhd1*, Emilin2, Fn1, Fndc1, Galnt16, Gap43, Heg1, Ifi27l2a, Igfbp4, Igfbp6, *Il18*, Kcnk5, Limch1, Lrrn4cl, Mfap5, Nid1, Npr1, Opcml, Pde8a, Pi16, *Pla1a*, Plac8, Plat, Prkg2, Procr, Prss23, Ptgis, Pygl, Rab32, Rcan2, Sema3c, Sfrp2, Slc43a3, *Smpd3*, Stmn4, Tek, Thbd, Wnt10b, Wnt2 [60 total, *8*]\n",
    "\n",
    "* x2 - w/x: 3300005D01Rik, *Ackr4*, Ccbe1, Cdkn2c, *Col7a1*, Cyp26b1, *Entpd1*, *Fam180a*, Fxyd5, *Gldn*, Gpm6b, Hmcn1, Meox2, Rspo4, S100a4, Tcf4, Thy1, Twist2 [18 total, *5*]\n",
    "\n",
    "* x1 - x2 (no x/y): Adcy1, Adh7, *Ahrr*, Aldh3a1, Ankrd29, *Axin2*, Cd63, Cdh4, Cldn10, Col3a1, Cox6c, Cyp26b1, Fgfr4, Gda, Gpnmb, *Grem1*, H2-D1, Lum, *Miat*, Nbl1, Ndufa4l2, Nt5e, Nupr1, Osr2, Pla2g2e, Plpp1, Plpp3, Por, Ppic, Ppp1r14a, *Ppp2r2c*, Rspo1, S100a4, Scara5, Sulf2, Tgfbi, Tmem132c, Tnfrsf19, Zfp536 [39 total, *5*]\n",
    "\n",
    "* x1 - x/y (no x2): Aebp1, Aldh1l2, C1qtnf3, Cd34, Cdh4, Cgref1, Cldn10, Clec11a, Cpz, Creb3l3, Crip2, Csf1r, Ctsh, Ctsk, Cyp2f2, Cyp4b1, Dapk1, Dcn, Ddah1, Ecm2, Fam46c, Gas6, Il1r2, Lgi1, Mmp27, P4ha2, Pdgfrl, Rarres2, Rcn3, Sectm1a, Sema3b, Serpina3n, Tgfb2, Tgfbi, Zfp536 [35 total, *0*]\n",
    "* x2 - x1 - x/y (all 3): Cdh4, Cldn10, Tgfbi, *Zfp536* [4 total, *1*]\n",
    "\n",
    "* x/y - y3: Abcc9, Art3, Aspn, C1qtnf3, Cilp, Col12a1, Fzd1, Gas6, Ghr, Gpx3, Igf1, Lbp, Lepr, Lgr5, Lox, Nfib, Nkain4, Ogn, Pcolce, Pcsk5, Pdgfrl, Plxna4, Rerg, Serpinf1, Sfrp1, Slit2, Tgfbi, Tgm2 [28 total, *0*]\n",
    "* x/y - y1: Agtr1a, Akap12, Akr1cl, Ctsh, Cxcl12, Dpep1, Dpt, Efemp1, Gpx3, Hpgd, *Lbp*, Lepr, Ly6a, Ly6c1, Mgst1, Nfib, Ogn, Pcsk5, Penk, *Pltp*, Slit2, Thbs3, Tnxb [23 total, *2]\n",
    "\n",
    "* y2 - y3: Abca8a, Adamtsl2, *Angptl1*, Aspn, Bgn, Cd9, *Cilp*, Cxcl14, Cygb, Dpysl3, F3, Fbln2, Fxyd6, Gfra1, Gpx3, Igf1, Igfbp7, Itm2a, Jam2, Lox, Ltbp4, Mgp, Mn1, Myoc, Nfib, Pcolce, Podn, Rbp1, Serpinf1, Sfrp1, Smoc2, Tgfbi [32 total, *2*]\n",
    "* y2 - y1: Adam19, Bicc1, Bmper, Col14a1, Col4a1, Col4a2, *Col5a3*, Col6a3, Col6a6, Cygb, Dpt, Fap, Fzd4, Ggt5, Gldc, Gpx3, Ifi205, Itih5, Itm2a, Lpl, Lsp1, Mfap5, Nfib, Slfn5, Srpx2, Thbs3, Tmeff2 [27 total, *0*]\n",
    "* y2 - y4: Adamtsl3, Adcyap1r1, Angpt1, B3gnt8, Bicc1, Bmper, Col4a1, Col4a2, Col4a4, Cp, Cst3, Cxcl14, Cygb, Entpd2, F3, Fxyd6, Fzd4, G0s2, Ggt5, Ifi203, Igfbp7, Il17d, *Inmt*, Lmo2, Loxl3, Lpl, Mgp, Msx1, Nfib, *Nmb*, Ntf3, P2ry1, P2ry14, Rbp1, Sept4, Slfn5, Sparcl1, Steap4, Tmeff2, Vtn [40 total, *2*]\n",
    "* y2 - y5: *Abca8a*, Bicc1, Col15a1, Col5a3, *Col8a1*, Cp, Entpd2, F3, *Fap*, Gfra1, Hspg2, Islr, Itm2a, *Meox1*, Mgp, Myoc, Nid1, P2ry14, Plxdc2, Smoc2, Sparcl1, Steap4, Tgfbi, Thbs1, Thbs4, Vtn [26 total, *3*]\n",
    "\n",
    "* y5 - y2 (no y4): *Abca8a*, Bicc1, Col15a1, Col5a3, *Col8a1*, Cp, Entpd2, F3, *Fap*, Gfra1, Hspg2, Islr, Itm2a, *Meox1*, Mgp, Myoc, Nid1, P2ry14, Plxdc2, Smoc2, Sparcl1, Steap4, Tgfbi, Thbs1, Thbs4, Vtn [26 total, *3*]\n",
    "* y5 - y4 (no y2): Abi3bp, Bicc1, C1s1, Ccl11, Cp, Cpe, Ebf2, *Ecm1*, Entpd2, F3, Gsn, Igfbp4, Il33, Mgp, Nr2f2, Ntrk2, P2ry14, Phlda1, Sparcl1, *Steap4*, Sult5a1, Tshz2, Vit, Vtn [24 total, *1*]\n",
    "* y5 - y2 - y4 (all 3): Bicc1, Cp, Entpd2, F3, Mgp, P2ry14, Sparcl1, *Steap4*, *Vtn* [9 total, *2*]\n",
    "\n",
    "* v1 - y5 (no y4): Abi3bp, Cpe, Cspg4, Ebf2, *Gfra1*, *Hmgcs2*, Il33, *Lmo4*, Mgp, Nr2f2, Sox9, *Sult5a1* [12 total, *4*]\n",
    "* v1 - y4 (no y5): Abi3bp, Acp5, Adamtsl3, B3gnt8, Bnc2, Cav1, *Ccdc3*, Cpe, Cttnbp2, Ebf2, Fxyd6, Il33, *Ildr2*, Mgp, *Ndrg2*, Nfib, Nr2f2, Pear1, Rgs16, Sult5a1 [20 total, *3*]\n",
    "* v1 - y5 - y4 (all 3): Abi3bp, Acp5, Adamtsl3, B3gnt8, Bnc2, Cav1, Ccdc3, Cpe, Cttnbp2, Ebf2, Fxyd6, Il33, Ildr2, Mgp, Ndrg2, Nfib, *Nr2f2*, Pear1, Rgs16, Sult5a1 [20 total, *X*]\n",
    "\n",
    "* w1 - w2: 6330403K07Rik, Acot1, *Alpl*, Alx4, Aplp1, Apoe, *Bambi*, Bcl2, *Bmp4*, C1qtnf4, *Cd24a*, Cebpa, Chodl, Chst8, Col23a1, Crabp1, *Crabp2*, Cttnbp2, Daam2, Ebf1, *Edn3*, Enpp2, Eva1a, Fgf10, Fgf7, Fgfr1, Fgfr2, Frem1, Gldn, Gnai1, Gng2, H2-Ab1, H3f3b, Heph, Hey2, *Hhip*, *Inhba*, *Kctd1*, Lamc3, Ldhb, *Lef1*, Lrrtm3, Ltbp1, Mdk, Mir155hg, *Mpped2*, Mpzl1, Ncam1, *Ndnf*, Ndp, Nkd2, *Notum*, Ntng1, Pappa, *Pappa2*, *Prdm1*, *Prlr*, Ptk7, Ptma, Ptprz1, Rab20, Rspo1, Rspo2, Rspo3, Runx3, *Scube3*, Serpine2, Slc16a2, Snhg11, Sostdc1, Spon1, Ss18, Ssbp2, Tceal3, *Tfap2c*, Tle2, Tmem132c, Tmem176a, Tmem176b, Trpm3, Trps1, Vcan, Wif1, a [84 total, *18*]\n",
    "* w3 - w4: Acan, *Acta2*, Actg2, Actn1, *Adamts18*, Bcl11b, *Bok*, Cd200, Cdc42ep3, Clic5, Cnn2, Col11a1, Csrp1, *Egfl6*, Etl4, Fam101b, Fam65b, Fry, Igfbp4, Itih5, Lmo4, *Lrrc15*, Mef2c, Mgst3, Myl9, *Mylk*, Nav2, Nradd, Nrtn, *Ntrk3*, Palld, Pard6g, Pawr, Phldb2, *Tagln*, Tcf7l2, Tmem119, Tnnt1, Tpbg, Tpd52l1, Tpm1, Tpm2, Vcl [43 total, *8*]\n",
    "\n",
    "\n",
    "* w3 - w/x - x1: 0\n",
    "* w3 - w/x - x2: 0\n",
    "\n",
    "* x2 - w4: *Igfbp2*, *Mme*, *Sema3a* [3 total, *3*]\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2022-05-03T09:52:46.091273Z",
     "iopub.status.busy": "2022-05-03T09:52:46.090756Z",
     "iopub.status.idle": "2022-05-03T09:52:46.125286Z",
     "shell.execute_reply": "2022-05-03T09:52:46.124363Z",
     "shell.execute_reply.started": "2022-05-03T09:52:46.091211Z"
    },
    "tags": []
   },
   "source": [
    "### Mouse-human comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "plot_best_N(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, N_min=10, N_max=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "N_human_mouse = 250\n",
    "df_jaccard_mouse_human = get_df_overlap(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, N=N_human_mouse)\n",
    "plot_heatmap(df_jaccard_mouse_human, dict_colors_mouse, dict_colors_human, figsize=(12,7), ticklabelsize=15, cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# sorted\n",
    "df_jaccard_mouse_human_sorted = df_jaccard_mouse_human.loc[['A2', 'A3', 'A1', 'A4', 'B1', 'B3', 'B2', 'B4', 'C2', 'C3', 'C1', 'C5', 'D1', 'D2', 'E1'], \n",
    "                                                           ['x2', 'x1', 'x/y', 'y3', 'y1', 'z2', 'z1', 'y4', 'y2', 'w/x', 'w4', 'w3', 'w1', 'w2', 'y5', 'v1', 'w5']]\n",
    "\n",
    "df_jaccard_mouse_human_sorted[df_jaccard_mouse_human_sorted < 0.05] = np.nan\n",
    "\n",
    "plot_heatmap(df_jaccard_mouse_human_sorted, dict_colors_mouse, dict_colors_human, figsize=(12,7), ticklabelsize=15, annot=True, cmap='Blues')\n",
    "\n",
    "plot_heatmap(df_jaccard_mouse_human_sorted, dict_colors_mouse, dict_colors_human, figsize=(12,7), ticklabelsize=15, annot=False, cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Case: mouse has two clusters to be considered\n",
    "cluster_1, cluster_2, cluster_2a = 'E1', 'w4', 'w3'\n",
    "\n",
    "list_genes_int = print_common_genes(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2, translate=True, N=250)\n",
    "\n",
    "list_genes_inta = print_common_genes(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2a, translate=True, N=250)\n",
    "\n",
    "list_genes_int = sorted(set(list_genes_int + list_genes_inta))\n",
    "print(', '.join(list_genes_int), f'[A, **X**, *Y*, ***Z***] - {len(list_genes_int)}')\n",
    "\n",
    "aaa_human = df_human_genes_codes.loc[[i for i in list_genes_int if i in df_human_genes_codes.index]]\n",
    "aaa_human = aaa_human.loc[np.array([cluster_1 in i for i in aaa_human['code'].values])]\n",
    "aaa_human.reset_index(inplace=True)\n",
    "\n",
    "aaa_mouse = df_mouse_genes_codes.loc[[i for i in [dict_human_mouse[j] for j in list_genes_int] if i in df_mouse_genes_codes.index]]\n",
    "aaa_mouse = aaa_mouse.loc[np.array([cluster_2 in i for i in aaa_mouse['code'].values]) | np.array([cluster_2a in i for i in aaa_mouse['code'].values]) | np.array([cluster_2[0] in i for i in aaa_mouse['code'].values])]\n",
    "aaa_mouse.reset_index(inplace=True)\n",
    "\n",
    "\n",
    "pd.concat([aaa_human, aaa_mouse], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Case: human has two clusters to be considered\n",
    "\n",
    "cluster_1, cluster_1a, cluster_2 = 'D1', 'D2', 'y4'\n",
    "# cluster_not = 'y5'\n",
    "\n",
    "list_genes_int = print_common_genes(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2, translate=True, N=250)\n",
    "\n",
    "list_genes_inta = print_common_genes(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1a, cluster_2, translate=True, N=250)\n",
    "\n",
    "list_genes_int = sorted(set(list_genes_int + list_genes_inta))\n",
    "print(', '.join(list_genes_int), f'[A, **X**, *Y*, ***Z***] - {len(list_genes_int)}')\n",
    "\n",
    "aaa_human = df_human_genes_codes.loc[[i for i in list_genes_int if i in df_human_genes_codes.index]]\n",
    "aaa_human = aaa_human.loc[np.array([cluster_1 in i for i in aaa_human['code'].values]) | np.array([cluster_1a in i for i in aaa_human['code'].values]) | np.array([cluster_1[0] in i for i in aaa_human['code'].values])]\n",
    "aaa_human.reset_index(inplace=True)\n",
    "\n",
    "aaa_mouse = df_mouse_genes_codes.loc[[i for i in [dict_human_mouse[j] for j in list_genes_int] if i in df_mouse_genes_codes.index]]\n",
    "aaa_mouse = aaa_mouse.loc[np.array([cluster_2 in i for i in aaa_mouse['code'].values]) | np.array([cluster_2[0] in i for i in aaa_mouse['code'].values])]\n",
    "aaa_mouse.reset_index(inplace=True)\n",
    "\n",
    "\n",
    "pd.concat([aaa_human, aaa_mouse], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "cluster_1, cluster_2 = 'E1', 'v1'\n",
    "# cluster_not = 'y5'\n",
    "\n",
    "list_genes_int = print_common_genes(dict_make_gene_scoring_cluster_robust_human, dict_make_gene_scoring_cluster_robust_mouse, \n",
    "                   cluster_1, cluster_2, translate=True, N=250)\n",
    "print(', '.join(list_genes_int), f'[A, **X**, *Y*, ***Z***] - {len(list_genes_int)}')\n",
    "\n",
    "aaa_human = df_human_genes_codes.loc[[i for i in list_genes_int if i in df_human_genes_codes.index]]\n",
    "aaa_human = aaa_human.loc[np.array([cluster_1 in i for i in aaa_human['code'].values])]\n",
    "aaa_human.reset_index(inplace=True)\n",
    "\n",
    "aaa_mouse = df_mouse_genes_codes.loc[[i for i in [dict_human_mouse[j] for j in list_genes_int] if i in df_mouse_genes_codes.index]]\n",
    "aaa_mouse = aaa_mouse.loc[np.array([cluster_2 in i for i in aaa_mouse['code'].values]) | np.array([cluster_2[0] in i for i in aaa_mouse['code'].values])]\n",
    "aaa_mouse.reset_index(inplace=True)\n",
    "\n",
    "\n",
    "pd.concat([aaa_human, aaa_mouse], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comments on human-mouse\n",
    "\n",
    "\n",
    "Representation of genes: MARKER, **MARKER IN HUMAN**, *MARKER IN MOUSE*, ***MARKER IN HUMAN AND MOUSE***  [1, **3**, *6*, ***3***]  -  13\n",
    "[MIRAR SI EL SOLAPAMIENTO ES MUCHO O POCO EN GENERAL]\n",
    "* A2 - x2: CD109, COL3A1, COL5A1, LOXL2, NFATC2, NOTUM, PAM, RSPO3, SCARF2, TCF4, ZNF608, **COL13A1**, **CYP26B1**, **DAAM2**, **ISM1**, **NKD1**, **NKD2**, **PAPPA**, **PREX1**, **PTGS1**, **TGFBI**, **THBD**, **THSD4**, *C10orf105*, *CAV1*, *CCBE1*, *EMX2*, *ENHO*, *LSAMP*, *MAMDC2*,  *SPRY1*, *TWIST2*, *WNT5A*, ***AHRR***, ***AXIN2***, ***CD9***, ***COL7A1***, ***F13A1***, ***GREM2***, ***IGFBP2***, ***KCNK2***, ***PTK7***, ***PTPRE***, ***RSPO1***, ***SMIM3***, ***STC1***, ***TNFRSF19*** [11, **12**, *10*, ***14***] - 47\n",
    "\n",
    "##### Esto es curioso porque A2 vemos que también comparte genes con C2 y derivados, así que este fenómeno es comun en human-human y human-mouse (pero no tan evidente en mouse-mouse)!!\n",
    "* A2 - w1: C4orf48, EMB, ETV1, PAPPA, PLCB1, SFRP2, WIF1, WNT5A, **APCDD1**, **AXIN2**, **FGFR2**, **NKD1**, **RAMP3**, **RGS2**, **SMIM3**, *LAMC3*, *NOTUM*, *NRN1*, *RSPO1*, *SPON1*, *TFAP2C*, ***APELA***, ***COL13A1***, ***COL23A1***, ***DAAM2***, ***F5***, ***NKD2***, ****PTK7***, ***RSPO3*** [8, **7**, *6*, ***8***] - 29\n",
    "\n",
    "* A1 - x1: C1QTNF3, CD34, CHPF, CLEC3B, COL1A1, COL1A2, COL3A1, CREB5, CTSK, DCN, ELN, FKBP9, HTRA1, IGFBP5, ITGBL1, KDELR3, LOXL4, MFAP4, MMP14, P4HA2, PCOLCE2, PDGFRL, PLPP1, RCN3, SLC38A10, SPARC, TSPAN4, VKORC1, **AEBP1**, **ANGPTL1**, **CADM3**, **GDF15**, **GPNMB**, **HPGD**, **MMP2**, **PODN**, **SCARA5**, **TIMP2**, **TNXB**, *ADAMTS2*, *PPIC*, ***CGREF1***, ***CPZ***, ***CYBRD1***, ***CYP4B1***, ***MMP27***, ***SEMA3B*** [28, **11**, *2*, ***6***] - 47\n",
    "\n",
    "* A1 - x/y: ADAMTSL1, ANGPTL1, BASP1, CCDC80, CD34, CILP, COL12A1, COL1A1, COL1A2, CTSB, CTSK, DCN, FKBP9, IGSF10, ITGBL1, NUCB2, OGN, P4HA2, MGST1, PCOLCE2, PPIB, RCN3, SERF2, SERPINF1, VKORC1,  **ABCC9**, **ADGRD1**, **AEBP1**, **AGTR1**, **CADM3**, **CYP4B1**, **GALNT15**, **HPGD**, **LOX**, **PCOLCE**, **PDGFRL**, **SEMA3B**, **SMOC2**, **SVEP1**, **THBS2**, **THBS3**, **TNXB**, *C1QTNF3*, *GPX3*, *GREB1L*, *PLTP*, ***CGREF1***, ***CPZ***, ***LGR5***, ***MMP27*** [25, **17**, *4*, ***4***] - 50\n",
    "\n",
    "* A1 - z2: CREB5, ECM1, FBN1, IGFBP5, IGFBP6, METRNL, PAMR1, PLXDC2, SEMA3C, TIMP3, VGLL3, **CHRDL1**, **CLEC3B**, **MFAP5**, **QPCT**, **SCARA5**, **TIMP2**, *ADAMTSL4*, *BASP1*, *CD248*, *CD34*, *COL14A1*, *DBN1*, *EMILIN2*, *LRRN4CL*, *PTGIS*, *SFRP2*, *UCHL1*, ***ACKR3***, ***ADGRD1***, ***CD55***, ***DPP4***, ***ISLR***, ***LIMS2***, ***MTCL1***, ***NPR1***, ***PI16***, ***PRKG2***, ***SEMA3E***, ***TUBB4A*** [11, **6**, *11*, ***12***] - 40\n",
    "\n",
    "* A1 - y3: ANGPTL1, COL12A1, COL14A1, CRYAB, CYB5R3, DHRS3, ELN, FGL2, GAS1, GPX3, IGFBP6, ITGBL1, MFAP4, OGN, PRELP, SERPINF1, SFRP2, TIMP3, **CLU**, **FBLN1**, **FBLN2**, **GALNT15**, **LGR5**, **LOX**, **CD151**, **PCOLCE**, **PDGFRL**, *C1QTNF3*, *CILP*, *MGP*, *PTGIS*, ***ABCC9***, ***OMD***, ***PODN***, ***SMOC2*** [18, **9**, *4*, ***4***] - 35\n",
    "\n",
    "\n",
    "\n",
    "* A4 - z1: ADAMTS5, FSTL1, GFPT2, MGLL, RAMP2, SDK1, TIMP2, TNFAIP6, VASN, ZYX, **ACE**, **CLEC3B**, **GPX3**, **IGFBP6**, **LOXL1**, **MFAP5**, **PTGIS**, **RAB32**, **SCARA5**, *ACKR3*, *ADGRD1*, *AXL*, *CHRDL1*, *FLNC*, *GAP43*, *HAS1*, *HEG1*, *PROCR*, *PRSS23*, *UGDH*, ***AIF1L***, ***CD248***, ***CD55***, ***DBN1***, ***DPP4***, ***EMILIN2***, ***NPR1***, ***SEMA3C***, ***SEMA3E***, ***SFRP4***, ***WNT2*** [10, **9**, *11*, ***11***] - 41\n",
    "\n",
    "* A4 - z2: BMP7, CD34, DDAH2, GFPT2, IGFBP5, MGLL, NHSL1, PCSK6, PPP1R14B, PXN, RAMP2, SCARA3, SDK1, SMURF2, TRIO, ZNF385A, ZYX, **ACE**, **ACKR3**, **CLEC3B**, **DBN1**, **FBN1**, **IGFBP6**, **LOXL1**, **MFAP5**, **SCARA5**, **TIMP2**, **TIMP3**, *ACKR2*, *ADAMTSL4*, *ADGRG2*, *AXL*, *CHRDL1*, *DACT2*, *GAP43*, *HEG1*, *PROCR*, *PRSS23*, *UGDH*, ***ADGRD1***, ***AIF1L***, ***CD248***, ***CD55***, ***DPP4***, ***EMILIN2***, ***ISLR***, ***LIMS2***, ***NPR1***, **PAMR1**, ***PTGIS***, ***RAB32***, ***SEMA3C***, ***SEMA3E***, ***SFRP4***, ***WNT2*** [17, **11**, *11*, ***16***] - 55\n",
    "\n",
    "* A4 - y1: ADAMTS5, ADGRD1, AHNAK2, APBB1IP, AXL, CAPG, CDH13, FAP, FSTL1, GFPT2, GLIPR2, HEG1, LGR4, LSP1, MEDAG, MGST1, MMP2, PRSS23, RAMP2, TAGLN2, THBS3, TMSB10, TNFAIP6, UGDH, VCAN, ZNF385A, **CD248**, **CD55**, **CLEC3B**, **EMILIN2**, **FBN1**, **GPX3**, **IGFBP6**, **LOXL1**, **MFAP5**, **TNXB**, **TPPP3**, *HMCN2*, *NOVA1*, *PLTP*, *SSC5D*, ***ACE***, ***CTHRC1*** [26, **11**, *4*, ***2***] - 43\n",
    "\n",
    "\n",
    "* z2 - A1: CREB5, ECM1, FBN1, IGFBP5, IGFBP6, METRNL, PAMR1, PLXDC2, SEMA3C, TIMP3, VGLL3, **CHRDL1**, **CLEC3B**, **MFAP5**, **QPCT**, **SCARA5**, **TIMP2**, *ADAMTSL4*, *BASP1*, *CD248*, *CD34*, *COL14A1*, *DBN1*, *EMILIN2*, *LRRN4CL*, *PTGIS*, *SFRP2*, *UCHL1*, ***ACKR3***, ***ADGRD1***, ***CD55***, ***DPP4***, ***ISLR***, ***LIMS2***, ***MTCL1***, ***NPR1***, ***PI16***, ***PRKG2***, ***SEMA3E***, ***TUBB4A*** [11, **6**, *11*, ***12***] - 40\n",
    "\n",
    "* z2 - A4: BMP7, CD34, DDAH2, GFPT2, IGFBP5, MGLL, NHSL1, PCSK6, PPP1R14B, PXN, RAMP2, SCARA3, SDK1, SMURF2, TRIO, ZNF385A, ZYX, **ACE**, **ACKR3**, **CLEC3B**, **DBN1**, **FBN1**, **IGFBP6**, **LOXL1**, **MFAP5**, **PAMR1**, **SCARA5**, **TIMP2**, **TIMP3**, *ACKR2*, *ADAMTSL4*, *ADGRG2*, *AXL*, *CHRDL1*, *DACT2*, *GAP43*, *HEG1*, *PROCR*, *PRSS23*, *UGDH*, ***ADGRD1***, ***AIF1L***, ***CD248***, ***CD55***, ***DPP4***, ***EMILIN2***, ***ISLR***, ***LIMS2***, ***NPR1***, ***PTGIS***, ***RAB32***, ***SEMA3C***, ***SEMA3E***, ***SFRP4***, ***WNT2*** [17, **12**, *11*, ***15***] - 55\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "* z1 - A4: ADAMTS5, FSTL1, GFPT2, MGLL, RAMP2, SDK1, TIMP2, TNFAIP6, VASN, ZYX, **ACE**, **CLEC3B**, **GPX3**, **IGFBP6**, **LOXL1**, **MFAP5**, **PTGIS**, **RAB32**, **SCARA5**, *ACKR3*, *ADGRD1*, *AXL*, *CHRDL1*, *FLNC*, *GAP43*, *HAS1*, *HEG1*, *PROCR*, *PRSS23*, *UGDH*, ***AIF1L***, ***CD248***, ***CD55***, ***DBN1***, ***DPP4***, ***EMILIN2***, ***NPR1***, ***SEMA3C***, ***SEMA3E***, ***SFRP4***, ***WNT2*** [10, **9**, *11*, ***11***] - 41\n",
    "* z1 - B1: ADAMTS1, BCL3, C3, CSRNP1, CSRP2, GFPT2, HMOX1, IFI16, MYC, NFE2L2, NFKBIZ, NOCT, PHLDA1, PIM1, PLSCR1, PNP, SOD2, **ERRFI1**, **KDM6B**, **MAFF**, **NFKB1**, **NFKBIA**, **NR4A3**, **REL**, **TIPARP**, **TNFAIP3**, **TNFAIP6**, **UAP1**, **ZC3H12A**, *HAS2*, *PTGS2*, *PTX3*, *TNFAIP2*, ***CCL2***, ***CXCL2***, ***FOSL1***, ***GCH1***, ***IL6***   [17, **12**, *4*, ***5***] - 38\n",
    " \n",
    "\n",
    "\n",
    "* y4 - B2: ABCA8, ABI3BP, C1S, FRMD6, HGF, ID4, IL11RA, LDB2, NRP1, OSMR, P2RY14, PLCXD3, TRIM47, TSPAN11, VEGFA, **APOC1**, **C7**, **CXCL12**, **CYGB**, **GGT5**, **IGFBP3**, **IGFBP7**, *ADCYAP1R1*, *AVPR1A*, *COL4A4*, *NDRG2*, *PTCH2*, *RBP1*, *SNED1*, *TMEM176A*, *TMEM176B*, ***APOE***, ***C3***, ***IL33***, ***MGP***, ***NFIB***, ***SLCO2B1***, ***TNFSF13B*** [15, **7**, *9*, ***7***] - 38\n",
    "* y4 - B4: ARHGDIB, COL4A2, CYGB, DACT1, EGFR, EPHA3, EPS8, F3, FAM13A, FOXP1, GSN, HGF, IGFBP3, IGFBP7, ITM2B, MGST1, NGFR, NID2, OLFML2B, SERPING1, SRPX, TMEM176B, TNFSF13B, TSHZ2, **ABCA8**, **APOC1**, **APOD**, **C7**, **CXCL12**, **FGF7**, **FMO1**, **FZD4**, **GGT5**, **GPX3**, **HHIP**, **IGF1**, **NFIB**, **NTRK2**, **PODN**, **VIT**, **ZFHX4**, *ABCC9*, *ADAMTSL3*, *APOE*, *BMPER*, *BNC2*, *C3*, *COL4A4*, *FMO2*, *IL33*, *INMT*, *NR2F2*, NRP1, *PEAR1*, *PPL*, *PTCH2*, *SFRP4*, *TMEM176A*, ***CHRDL1***, ***GDF10***, ***MGP*** [24, **17**, *17*, ***3***] - 61\n",
    "\n",
    "\n",
    "\n",
    "##### Sensación: hay genes de y2-B4 que no están en negrita-cursiva porque son más de B que de B4. Entonces quizás y2 tenga una relación más immune, pero no con un cluster específico.\n",
    "* y2 - B4: ARPC1B, CAPN6, COL15A1, COL4A2, COL4A4, CRLF1, EGFR, EPS8, F3, FHL2, HIC1, HSPG2, HTRA3, IGFBP7, ITM2A, ITM2B, NID2, SPRY1, VWA1, **ABCA9**, **APOD**, **CYGB**, **FMO1**, **FZD4**, **GGT5**, **GPX3**, **GSN**, **IGF1**, **LSP1**, **MGP**, **MYOC**, **NFIB**, **PODN**, **TSHZ2**, **ZFHX4**, *ADAMTSL3*, *BMPER*, *INMT*, *PEAR1*, ***ABCA8*** [19, **16**, *4*, ***1***] - 40\n",
    "* y2 - A1: ANGPTL2, CAPG, GPX3, HTRA3, IL17D, ITIH5, OGN, PLXDC2, PRELP, PRG4, **FBLN2**, **ISLR**, **LOX**, **MFAP5**, **PCOLCE**, **PODN**, **SERPINF1**, **SLC29A1**, **SMOC2**, **SVEP1**, **THBS3**, **TNXB**, *CILP*, *COL14A1*, *FAP*, *MGP*, *SSC5D*, ***ANGPTL1*** [10, **11**, *5*, ***1***] - 28\n",
    "\n",
    "\n",
    "* B4 - y4: ARHGDIB, COL4A2, CYGB, DACT1, EGFR, EPHA3, EPS8, F3, FAM13A, FOXP1, GSN, HGF, IGFBP3, IGFBP7, ITM2B, MGST1, NGFR, NID2, OLFML2B, SERPING1, SRPX, TMEM176B, TNFSF13B, TSHZ2, **ABCA8**, **APOC1**, **APOD**, **C7**, **CXCL12**, **FGF7**, **FMO1**, **FZD4**, **GGT5**, **GPX3**, **HHIP**, **IGF1**, **NFIB**, **NTRK2**, **PODN**, **VIT**, **ZFHX4**, *ABCC9*, *ADAMTSL3*, *APOE*, *BMPER*, *BNC2*, *C3*, *COL4A4*, *FMO2*, *IL33*, *INMT*, *NR2F2*, NRP1, *PEAR1*, *PPL*, *PTCH2*, *SFRP4*, *TMEM176A*, ***CHRDL1***, ***GDF10***, ***MGP*** [24, **17**, *17*, ***3***] - 61\n",
    "* B4 - y2: ARPC1B, CAPN6, COL15A1, COL4A2, COL4A4, CRLF1, EGFR, EPS8, F3, FHL2, HIC1, HSPG2, HTRA3, IGFBP7, ITM2A, ITM2B, NID2, SPRY1, VWA1, **ABCA9**, **APOD**, **CYGB**, **FMO1**, **FZD4**, **GGT5**, **GPX3**, **GSN**, **IGF1**, **LSP1**, **MGP**, **MYOC**, **NFIB**, **PODN**, **TSHZ2**, **ZFHX4**, *ADAMTSL3*, *BMPER*, *INMT*, *PEAR1*, ***ABCA8*** [19, **16**, *4*, ***1***] - 40\n",
    "* B4 - y3: ARHGDIB, DEPTOR, DHRS3, EPHA3, F3, FHL2, GHR, GPC3, IGFBP3, IGFBP6, IGFBP7, KCNJ8, NPY1R, PEAR1, SFRP4, SLIT2, SUSD2, TIMP3, **ABCA8**, **APOD**, **CYGB**, **FGF7**, **FMO1**, **GPX3**, **GSN**, **IGF1**, **ITM2A**, **MGP**, **MYOC**, **NFIB**, **PHLDA3**, **PODN**, **TXNIP**, **ZFHX4**, *ABCC9*, *FMO2* [18, **16**, *2*, ***0***] - 36\n",
    "* B4 - y1: APBB1IP, ARPC1B, COL4A2, EBF1, EGFR, GPC3, HIC1, IGFBP3, LGALS3BP, LGMN, MEDAG, NFIB, NR1H3, PRSS23, SLIT2, TGFBR2, THY1, TMEM135, **CXCL12**, **CYGB**, **EFEMP1**, **FMO1**, **FZD4**, **GGT5**, **GPX3**, **IGFBP6**, **ITM2A**, **LSP1**, **MGST1**, **RARRES2**, **ZFHX4**, *BMPER*, *FABP4*, *NOVA1*, *PLAT*, ***PPARG*** [18, **13**, *4*, ***1***] - 36\n",
    "\n",
    "* C2 - w/x: ADAMTS9, CBFA2T3, COL11A1, EDNRA, ENHO, HTRA1, SLC48A1, SRPX, TBXA2R, TENM3, **CPNE5**, **CRABP1**, **FZD1**, **GPM6B**, **MEOX2**, **NCAM1**, **NFATC2**, **PLXDC1**, **PTH1R**, **RSPO4**, **SLC40A1**, **TBX15**, **TCF4**, **TRIB2**, **TRPS1**, *HS3ST6*, *KIF26B*, *MEGF6*, *NR2F1*, *TSHZ3*, ***CHST15***, ***COCH***, ***CYP1B1***, ***DKK2***, ***EMID1***, ***FIBIN***, ***FMOD***, ***MAFB***, ***MKX***, ***NRP2***, ***PTGFR***, ***TNMD*** [10, **15**, *5*, ***12***] - 42\n",
    "* C2 - w1|w2: CNTN1, CTTNBP2, PRLR, PTGER3, PTPRD, SFRP1, TSPAN7, **ATP1B1**, **BTBD11**, **CLEC14A**, **DKK2**, **EMB**, **GPM6B**, **IGFBP5**, **MEIS2**, **NCAM1**, **SPARCL1**, *ALX4*, *CRABP2*, *DUSP10*, *LEF1*, *NDP*, *PDE1A*, *PTK7*, *RSPO4*, *RUNX3*, *SDC1*, *WNT5A*, ***CRABP1***, ***DAAM2***, ***NDNF***, ***NOTUM***, ***TRPM3***, ***TRPS1*** [7, **10**, *11*, ***6***] - 34\n",
    "* C2 - y3: CALM2, MDFIC, OGN, PDE4B, PMEPA1, PRELP, RERG, SERTAD4, SUSD2, **EMB**, **FIBIN**, **FMOD**, **FZD1**, **KCNAB1**,  **PTPRD**, **PTPRK**, **TBX15**, *CAVIN2*, *F3*, *FGF9*, *LTBP4*, *PAQR6*, *SFRP1*, ***ASPN***, ***CADM1*** [9, **8**, *6*, ***2***] - 25\n",
    "\n",
    "\n",
    "* C3 - w/x: ADAMTS9, AQP1, CDH11, CHCHD10, CNN2, COL11A1, COL16A1, CRABP1, DKK3, EDNRA, EGFLAM, EMP2, FIBIN, FZD1, GPM6B, HMCN1, MEF2C, MFAP4, MMP11, MXRA8, PLXDC1, PTMA, RSPO4, SESN3, SLC40A1, TBX15, TCF4, THY1, TMEM119, TMEM204, TPM2, TRPS1, **F2R**, **HTRA1**, **MFAP2**, **POSTN**, **TENM3**, *CPXM2*, *DKK2*, *EMID1*, *KIF26B*, *MEGF6*, *NR2F1*, *NREP*, *NRP2*, *RFLNB*, *TSHZ3*, ***COL7A1***, ***MAFB***, ***MMP16***, ***RASL11B*** [32, **5**, *10*, ***4***] - 51\n",
    "* C3 - w4: BGN, BHLHE41, CALD1, CDH11, CNN2, EMP2, FABP5, ITGB1, ITGBL1, MGLL, MYH9, NPNT, PALLD, PPIC, PRELP, SORCS2, SPARC, **KIAA1217**, **LMO7**, **PMEPA1**, **SEMA5A**, *ACAN*, *ACTN1*, *ADAMTS9*, *AQP1*, *CGNL1*, *COL11A1*, *COL12A1*, *CPXM2*, *EDNRA*, *EGFLAM*, *LBH*, *MEF2C*, *MME*, *SLC40A1*,  *TAGLN*, *THBS4*, *TMEM119*, *TPM2*, ***COL7A1***, ***COL8A2***, ***LRRC15***, ***POSTN***, ***TENM3***   [17, **4**, *18*, ***5***] - 44\n",
    "* C3 - x1: BMP1, CHPF, COL3A1, COPZ2, GLT8D2, ITGBL1, LUM, MARVELD1, MFAP4, MMP14, MMP2, P4HA2, PYCR1, RCN3, RRBP1, SLC16A3, SPARC, SPHK1, SPON2, **BGN**, **C1QTNF6**, **COL5A1**, **COL5A2**, **ELN**, **HTRA1**, *ADAMTS2*, *AEBP1*, *COL16A1*, *COL1A1*, *COL1A2*, *EMID1*, *PPIC*, *PPP1R14A*, *SULF2*, ***LTBP2*** [19, **6**, *9*, ***1***] - 35\n",
    "\n",
    "* C1 - w4: ARHGAP28, ATP10A, CDH11, COL27A1, COL4A1, KIAA1217, KLF5, LMO7, MGLL, NPNT, NUAK1, NXN, PALLD, PMEPA1, POSTN, RNF152, SPARC, **MME**, **CNN2**, **SEMA5A**, *ADAMTS6*, *CD200*, *COL7A1*, *FOXD2*, *LMO4*, *LRRC15*, *NID2*, *NTRK3*, *PARD6G*, *PAWR*, *PRR5L*, *SATB2*, *TMEM119*, *TNMD*, *TPM2*, ***ACAN***, ***ACTA2***, ***ADAMTS18***, ***ADAMTS9***, ***BCL11B***, **CALD1**, ***CCND1***, ***COL11A1***, ***COL12A1***, ***COL8A2***, ***CPXM2***, ***EDNRA***, ***EDNRB***, ***EGFL6***, ***MEF2C***, ***RAMP1***, ***TAGLN***, ***TENM3*** [17, **3**, *15*, ***18***] - 53\n",
    "* C1 - w3: AFAP1L2, COL6A3, FNBP1L, FOXD2, JAG1, KIF26B, LMO4, MDK, MGLL, MICAL2, NPM1, PALLD, PARD6G, PAWR, PMEPA1, SHOX2, SOX18, SOX4, TMEM119, TNS3, TPM2, **ACAN**, **ACTA2**, **ADAMTS18**, **ALX4**, **BCL11B**, **CALD1**, **CD200**, **CDH11**, **CNN2**, **COL11A1**, **DKK3**, **EDNRA**, **EVA1A**, **F2R**, **KIAA1217**, **LAMC3**, **MEF2C**, **PTCH1**, **ROBO2**, **RUNX2**, **STMN1**, *LRRC15*, *MMP11*, *NTRK3*, *TAGLN*, ***CCDC3***, ***EGFL6*** [21, **21**, *4*, ***2***] - 48\n",
    "\n",
    "\n",
    "* C5 - w3: CFL1, COL11A1, DKK3, EDNRA, EEF1A1, F2R, PFN2, PMEPA1, PTMA, SMS, SOX4, STMN1, **ALX4**, **CXCR4**, **HEY2**, **JAG1**, **LMO4**, **MARCKSL1**, **PRDM1**, **PREX2**, **PTCH1**, **SOX18**, *KIF26B*, *LAMC3*, ***CDH11***, ***ROBO2*** [12, **10**, *2*, ***2***] - 26\n",
    "* C5 - w1|w2: C4orf48, FST, LAMC3, MARCKSL1, PLK2, PTCH1, PTMA, RAB34, SPARCL1, **CXCR4**, **IGFBP3**, **PDE3A**, **PRDM1**, **PREX2**, **ROBO1**, *EDN3*, *HEY2*, *SDC1*, *SLC26A7*, *SPON1*, *TRPS1*, ***ALX4***, ***BMP7***, ***INHBA***, ***LEF1***, ***PGM2L1***, ***SOX18***, ***TFAP2A***, ***WNT5A*** [9, **6**, *6*, ***8***] - 29\n",
    "\n",
    "\n",
    "\n",
    "* D1|D2 - v1: BHLHE40, CSPG4, CSRP1, DMD, DOCK9, ENDOD1, FRMD4B, JAG1, KTN1, MDFIC, MFAP5, NDRG1, PHLDA3, PLSCR4, PTCH1, SFRP1, STMN1, SYNE2, TLN2, TUBA4A, **ADAMTSL3**, **AKAP12**, **CAVIN2**, **CD200**, **CTNNAL1**, **DACT1**, **DUSP5**, **EFNA1**, **ETV1**, **IGFBP6**, **ISYNA1**, **MRAS**, **MTUS1**, **NR2F2**, **OLFML2A**, **PEAR1**, **PLEKHA4**, **SLC2A1**, **SORBS1**, **SOX9**, **STXBP6**, **TGFBI**, **TJP1**, **VIT**, *CCDC3*, *DDIT4*, *EZR*, *PERP*, *TPD52*, *TRIB2*, ***AQP3***, ***BNC2***, ***CAV1***, ***CAV2***, ***CLDN1***, ***EBF2***, ***EFNB1***, ***EGR3***, ***ETV4***, ***GAB1***, ***GPC1***, ***ITGA6***, ***ITGB4***, ***KLF5***, ***KRT19***, ***MTSS1***, ***NDRG2***, ***SBSPON***, ***SLC12A2***, ***TENM2*** [20, **24**, *6*, ***20***] - 70\n",
    "* D1|D2 - y5: ARL4A, CDKN2B, CNN3, CSPG4, GPC3, ITM2A, LUM, NID1, PHLDA1, SMOC2, SPARCL1, **ABCA8**, **ENTPD2**, **ETV1**, **TGFBI**, **TM4SF1**, ***APOD***, ***COL8A1***, ***EBF2***, ***FOXS1***, ***MATN2***, ***NR2F2***, ***P2RY14***, ***SOX9***, ***VIT*** [11, **5**, *0*, ***9***] - 25\n",
    "* D1|D2 - y4: CCDC3, COL4A2, CYGB, EPS8, FMO1, IGFBP7, ITM2B, MFAP5, NR2F1, PHLDA1, SPARCL1, WFDC1, **ABCA8**, **APOD**, **BNC2**, **CAV1**, **CYP1B1**, **DACT1**, **EBF2**, **EGFR**, **ENTPD2**, **GPC6**, **ITGA6**, **KLF5**, **NGFR**, **P2RY14**, **S100B**, **SCN7A**, **SFRP4**, *CHRDL1*, *LTBP4*, ***ADAMTSL3***, ***FMO2***, ***INMT***, ***MEOX2***, ***NDRG2***, ***NR2F2***, ***PEAR1***, ***VIT*** [12, **17**, *2*, ***8***] - 39\n",
    "\n",
    "\n",
    "* y5 - B4: ABCA8, F3, GPC3, HSPG2, IL33, ITM2A, LXN, MGP, RARRES1, **COL15A1**, **GSN**, **NTRK2**, *ABCA6*, *APOD*, *FOXS1*, *NR2F2*, *PHGDH*, *VIT*, *VWA1*, ***MYOC***, ***TSHZ2*** [9, **3**, *7*, ***2***] - 21\n",
    "* y5 - D1: ARL4A, CNN3, CSPG4, ITM2A, LUM, NID1, **EBF2***, **ENTPD2**, **ETV1**, **GPC3**, **PHLDA1**, **SPARCL1**, **TM4SF1**, *CDKN2B*, *MATN2*, ***ABCA8**, ***APOD***, ***COL8A1***, ***FOXS1***, ***NR2F2***, ***P2RY14***, ***SOX9***, ***TGFBI***, ***VIT*** [6, **7**, *2*, ***9***] - 24\n",
    "* y5 - D2: ABCA8, CNN3, EBF2, ETV1, ITM2A, NR2F2, SMOC2, TGFBI, TM4SF1, VIT, *FOXS1*, ***MATN2***, ***P2RY14*** [10, **0**, *1*, ***2***] - 13\n",
    "\n",
    "\n",
    "* v1 - D1: CAVIN2, CSPG4, CSRP1, CTNNAL1, DOCK9, FRMD4B, JAG1, MTUS1, NDRG1, PEAR1, PLEKHA4, PTCH1, STMN1, SYNE2, TLN2, TUBA4A, **CD200**, **DUSP5**, **EFNA1**, **EGR3**, **ETV1**, **MRAS**, **NR2F2**, **OLFML2A**, **SLC12A2**, **SOX9**, **TENM2***, **TGFBI**, **VIT**, *BNC2*, *EZR*, *ITGA6*, *ITGB4*, *KLF5*, *SBSPON*, *SLC2A1*, ***AKAP12***, ***CLDN1***, ***EBF2***, ***EFNB1***, ***ETV4***, **MTSS1***, ***NDRG2*** [16, **13**, *7*, ***7***] - 43\n",
    "* v1 - D2: BHLHE40, DMD, ENDOD1, ETV1, FRMD4B, JAG1, KTN1, MDFIC, MFAP5, NDRG1, NR2F2, PHLDA3, PLSCR4, PTCH1, SFRP1, STMN1, STXBP6, SYNE2, TLN2, **ADAMTSL3**, **AKAP12**, **CAVIN2**, **CSRP1**, **DACT1**, **DUSP5**, **IGFBP6**, **ISYNA1**, **MRAS**, **OLFML2A**, **PEAR1**, **PLEKHA4**, **SORBS1**, **TGFBI**, **TJP1**, **VIT**, *CCDC3*, *DDIT4*, *EZR*, *PERP*, *TPD52*, *TRIB2*, ***AQP3***, ***BNC2***, ***CAV1***, ***CAV2***, ***CLDN1***, ***DOCK9***, ***EBF2***, ***EFNB1***, ***GAB1***, ***GPC1***, ***ITGA6***, ***ITGB4***, ***KLF5***, ***KRT19***, ***MTSS1***, ***NDRG2***, ***SBSPON***, ***SLC2A1***, ***TENM2*** [19, **16**, *6*, ***19***] - 60\n",
    "\n",
    "\n",
    "* y4 - D1: APOD, BNC2, EBF2, EGFR, EPS8, FMO1, IGFBP7, ITGA6, KLF5, LTBP4, MEOX2, NDRG2, NGFR, NR2F2, PEAR1, S100B, SCN7A, VIT, **CYP1B1**, **P2RY14**, **PHLDA1**, **SPARCL1**, **WFDC1**, *ABCA8*, *CYGB*, *INMT*, *SFRP4*, ***ENTPD2***, ***FMO2*** [18, **5**, *4*, ***2***] - 29\n",
    "* y4 - D2: ABCA8, ADAMTSL3, CCDC3, COL4A2, EBF2, EGFR, CYP1B1, IGFBP7, ITM2B, LTBP4, MEOX2, MFAP5, NDRG2, NR2F1, NR2F2, P2RY14, SCN7A, VIT, **BNC2**, **CAV1**, **DACT1**, **GPC6**, **ITGA6**, **KLF5**, **NGFR**, **PEAR1**, *CHRDL1*, *SFRP4*, ***INMT*** [18, **8**, *2*, ***1***] - 29\n",
    "\n",
    "* E1 - y2: APOD, CAPG, COL15A1, FHL2, G0S2, GSN, IGF1, ITGA11, ITM2A, ITM2B, JUP, LAMA2, LPL, LSP1, MFAP5, SCN7A, SFRP1, SPRY1, TMEM204, **CMKLR1**, **MGP**, **PLEKHA6**, **RGMA**, *COL14A1*, *CRLF1*, *GPM6B*, *INMT*, *PEAR1*, *TGFBI*, *TSHZ2*, *VWA1*, ***MEOX1*** [19, **4**, *8*, ***1***] - 32\n",
    "\n",
    "* E1 - w3|w4: ACOT7, ALX4, BHLHE40, BHLHE41, C1QTNF7, CALD1, CCDC34, CDC42EP3, CDK6, DKK3, EFNB1, EMP2, EVA1A, FHL2, HES1, HEY1, ITGB1, JUP, KLF5, LAMC3, LMO7, MAF, MFGE8, MSI2, RASGRP2, RGCC, RUNX3, SEMA3G, SEMA5A, SHOX2, TBX18, **EGR2**, **KIAA1217**, **OLFML2A**, **RAMP1**, **TCF7L2**, *AQP1*, *ARHGDIB*, *CDH11*, *COL8A2*, *CSRP1*, *EDNRA*, *F2R*, *FOXD2*, *PARD6G*, *RUNX2*, *STMN2*, ***IGFBP2***, ***NTRK3*** [31, **5**, *11*, ***2***] - 49"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comments on human-mouse\n",
    "Representation of genes: MARKER, **MARKER IN HUMAN**, *MARKER IN MOUSE*, ***MARKER IN HUMAN AND MOUSE***  [1, **3**, *6*, ***3***]  -  13\n",
    "\n",
    "Of note, the selection of genes is based on the [LOC (fb, UMAP)] column from [this table](https://docs.google.com/spreadsheets/d/1lfI6sgjEyg37BGL7VRMfW7KgwGKwX5QrCtnKYk1DXY4/edit?usp=sharing). To make this table I based on the UMAPs from 4H and 4M notabooks. The notation is X ~ Y if two clusters or axes are equally *relevant*, and X > Y if X is *more relevant* than Y. **This notation is subjective** but to some extent necessary.\n",
    "\n",
    "The logic for a marker on the following list to be chosen **is also subjective** and is based in two factors: either (1) the marker is exclusive of the cluster it is said to be represented (e.g. if it says is of cluster b5 and the UMAP show it is only expressed in cluster b5 -- b5 -- or it is expressed in b5 preferentially -- b5 > X --) or (2) it is a marker of that cluster, among others (b5 ~ X or b ~ X > Y).\n",
    "    \n",
    "---\n",
    "\n",
    "* A4: both a2 and a1 show a great overlap with A4. Looking at the markers, a2 has 2 fully overlapping markers (mouse/human) and a1 none, but this is not very relevant. As good markers I would choose IGFBP6, SEMA3E and PCOLCE2.\n",
    "    * A4 - a1:  ADGRD1, FLNC, VASN, **ACE**, *HAS1*, *METRNL*, *TNFAIP6*, *UGDH*  -  [3, **1**, *4*, ***0***]  -  8\n",
    "    * A4 - a2: ADAMTSL4,  LIMS2, **CLEC3B**, **IGFBP5**, **ISLR**, **PAMR1**, **PTGIS**, **SFRP4**, *ADGRG2*, ***ACKR3***,  ***RAB32***  -  [2, **6**, *1*, ***2***]  -  11\n",
    "    * A4 - a1 - a2: CD55, CHRDL1, CREB5, DPP4, GAP43, HEG1, PRSS23, SCARA5, TIMP2, **CD248**, **DBN1**, **EMILIN2**, **SEMA3C**, *ALDH1A3*, *BASP1*, *FNDC1*, *GFPT2*, *MFAP5*, *NPR1*, ***AIF1L***, ***IGFBP6***, ***PCOLCE2***, ***SEMA3E***, ***WNT2***  -  [9, **4**, *6*, ***5***]  -  24\n",
    "    \n",
    "    \n",
    "* A1: in general there is a good gene overlap within all groups. Maybe b/c is the less overlapping group. Curiously, the \"good markers\" between A1 and a2, are different to the ones between A1 and b3 or A1 and c1. This migh t be that each mouse cluster has independent functions, which are shared with A1.\n",
    "    * A1 - a2 [endo/immune, A4-like]: DBN1, LIMS2, PAMR1, PTGIS, **CD55**, **CHRDL1**, **CLEC3B**, **ECM1**, **ISLR**, **SCARA5**, **SFRP2**, **TIMP2**, **TUBB4A**, *EMILIN2*, *IGFBP6*, *SEMA3C*, *UCHL1* , ***ACKR3***, ***CD248***, ***DPP4***, ***MFAP5***, ***PI16***, ***PRKG2***  -  [4, **9**, *4*, ***6***]  -  23\n",
    "    \n",
    "    * A1 - b3 [B-like]: C1QTNF3, FGL2, GPX3, **CD151**, **FBLN1**, **FBLN2**, **PCOLCE**, **SERPINF1**, **SMOC2**, *COL12A1*, *ELN*, *MFAP4*, *MGP*, *PAM*, *PTGIS*, ***ABCC9***, **ANGPTL1**, ***CCN5***, ***CLU***, ***LGR5***, ***LOX***, ***OMD***, ***PDGFRL***, ***PODN***, ***SFRP2***   -  [3, **7**, *6*, ***9***]  -  25\n",
    "    \n",
    "    * A1 - b/c: CCDC80, CD34, COL12A1, GALNT15, GPX3, MGST1, **AEBP1**, **ANGPTL1**, **CADM3**, **CPZ**, **CTSK**, **DCN**, **HPGD**, **PCOLCE**, **PDGFRL**, **SEMA3B**, **SERPINF1**, **THBS3**, **TNXB**, *C1QTNF3*, *CYP4B1*, *MMP27*, *PLTP*, ***ABCC9***, ***LGR5***, ***THBS2***  -  [6, **13**, *4*, ***3***]  -  26\n",
    "    \n",
    "    * A1 - c1 [ECM breakdown, C-like]: COL1A1, SPARC, **AEBP1**, **ANGPTL1**, **COL1A2**, **DCN**, **PDGFRL**, **SCARA5**, *C1QTNF3*, *ELN*, *ITGBL1*, *MFAP4*, *PCOLCE2*, ***CPZ***, ***CTSK***, ***CYBRD1***, ***HPGD***, ***MMP27***, ***SEMA3B***,  -  [2, **6**, *5*, ***6***]  -  19\n",
    "    \n",
    "    \n",
    "* A3: Considering it as a \"bridge\" population it is clear that there are not very good overlappping markers.\n",
    "    * A3 - b2: CD81, CD9, COL6A1, COL6A2, CP, ENPP3, FBLN2, LOX, MFAP5, SMOC2, **COL14A1**, **SERPINF1**, **SVEP1**, **THBS3**, *ISLR*, *PODN*, *SSC5D*, ***ITIH5***, ***THBS4***  -  [10, **4**, *3*, ***2***]  -  19\n",
    "    \n",
    "    * A3 - c1: AEBP1, COL1A1, COL1A2, COL5A1, COPZ2, DCN, HPGD, HTRA1, KDELR3, P4HA2, PDGFRL, RCN3, RSPO1, SERPINH1, SPARC, TSPAN4, **COL3A1**, *ADAMTS2*, *ALDH3A1*, *C1QTNF3*, *CGREF1*, *CTSK*, *CYBRD1*, *ECM2*, *MFAP4*, *PCOLCE2*, *PPIC*, *SEMA3B*,  ***ELN***, ***MMP27***  -  [16, **1**, *11*, ***2***]  -  30\n",
    "    \n",
    "    * A3 - b2 - c1: **ANGPTL1**, **BGN**  -  [0, **2**, *0*, ***0***]  -  2\n",
    "\n",
    "---\n",
    "    \n",
    "* A2\n",
    "    * A2 - c2: COL3A1, EMX2, LAMC3, PTK7, SCARF2, SMIM3, SPON1, TCF4, TGFBI, ZNF608, **COL23A1**, **CYP26B1**, **NKD1**, **NKD2**, **PREX1**, **RSPO1**, *AXIN2*, *CD9*, *COL7A1*, *HS3ST6*, *IGFBP2*, *MAMDC2*, ***AHRR***, ***F13A1***, ***GREM2***, ***KCNK2***, ***LSAMP***, ***PTPRE***, ***SPRY1***, ***STC1***, ***TNFRSF19***, ***TWIST2***   -  [10, **6**, *6*, ***10***]  -  32\n",
    "\n",
    "---\n",
    "\n",
    "* B1 and B3 vs b6 : Some of the genes (CCL2, CXCL2) are also expressed in a1. Also, it is likely that B1 is more related to b6 than B3, but this would also be somewhat expected because it is a bridge cluster.\n",
    "    * B1 - b6: ARID5A, ARL5B, BCL3, ETS2, GFPT2, PTGS2, TNFAIP2, **CEBPB**, **ERRFI1**, **FOSL1**, **GCH1**, **IER3**, **KDM6B**, **MAFF**, **NR4A3**, **PNRC1**, *CSRNP1*, *LIF*, ***ELL2***, ***IL6***, ***NFKB1***, ***TNFAIP6***  -  [7, **9**, *2*, ***4***]  -  22\n",
    "    * B3 - b6: ATF3, BTG2, CSF1, FOSB, GADD45B, OSMR, RARRES2, RNF122, **CXCL12**, **EGR1**, **JUNB**, **TMEM176A**, **TMEM176B**, **VCAM1**, **ZFP36**, ***CYP7B1***  -  [8, **0**, *7*, ***1***]  -  16\n",
    "    * B1 - B3 - b6: ADAMTS1, BTG1, CCNL1, MYC, NNMT, UGCG, **ADAMTS4**, **CCL2**, **CXCL2**, **ICAM1**, **IRF1**, **SOCS3**, **SOD2**, **TNFAIP3**, *NFKBIZ*, ***BIRC3***, ***NFKBIA***   -  [6, **8**, *1*, ***2***]  -  17\n",
    "\n",
    "---\n",
    "\n",
    "* B2 and B4 vs b4: there is not a clear relationship between b4 and B2/B4. \n",
    "    * B2 - b4: C1S, COL4A2, FRMD6, IL11RA, P2RY14, **IL33**, *AVPR1A*, *ID4*, *NDRG2*, *PTCH2*, *SNED1*, ***COL4A4***, ***TMEM176B***  -  [5, **1**, *5*, ***2***]  -  13\n",
    "    * B4 - b4: BMPER, FZD4, PPL, SERPING1, SRPX, **GPX3**, **GSN**, **NFIB**, **TSHZ2**, *ADAMTSL3*, *ADCYAP1R1*, *F3*, *VIT*, ***GDF10***, ***MGP***, ***NTRK2***  -  [5, **4**, *4*, ***3***]  -  16\n",
    "    * B2 - B4 - b4: C3, **GGT5**, *APOE*, *IGFBP7*, *NRP1*, *TNFSF13B*, ***C7***, ***CXCL12***, ***CYGB***, ***TMEM176A***  -  [1, **1**, *4*, ***4***]  -  10\n",
    "\n",
    "\n",
    "* B4 vs b1 and b4\n",
    "    * B4 - b1: EBF1, GPC3, LGALS3BP, NOVA1, **EFEMP1**, **FMO1**, **ITM2A**, **MGST1**, *CD36*, *FABP4*, ***C6***, ***PPARG***  -  [4, **4**, *2*, ***2***]  -  12\n",
    "    * B4 - b4: PPL, SERPING1, SRPX, **GSN**, **NFIB**, *ADAMTSL3*, *ADCYAP1R1*, *APOE*, *C3*, *C7*, *F3*, *NRP1*, *TMEM176A*, *TNFSF13B*, *VIT*, ***GDF10***, ***MGP***, ***NTRK2***, **TSHZ2**  -  [3, **2**, *10*, ***3***]  -  18\n",
    "    * B4 - b1 - b4: BMPER, FZD4, **GGT5**, **GPX3**, *IGFBP7*, ***CXCL12***, ***CYGB***  -  [2, **3**, *1*, ***2***]  -  8\n",
    "\n",
    "---\n",
    "Since we \"already know\" by Joost et al. 2020 that these populations belong to DS and DP, they are not THAT relevant to be studied as novel populations. However, considering the nice overlap between each pair (C1-d3, C5-d1/d2, C2-c/d), maybe doing a human-mouse pair to pair comparison would be nice.\n",
    "\n",
    "* C1\n",
    "    * C1 - d3: KIAA1217, MDK, MICAL2, STMN1, TNN, TNS3, **ALX4**, **EDNRA**, **EDNRB**, **LAMC3**, **PTCH1**, **ROBO2**, **RUNX2**, **TENM3**, *CD200*, *EGFL6*, *NTRK3*, *PALLD*, *TAGLN*, *TMEM119*, *TPM2*,  ***ADAMTS18***, ***BCL11B***, ***CDH11***, ***CNN2***, ***COL11A1***, ***F2R***, ***KIF26B***, ***MEF2C***  -  [6, **8**, *7*, ***8***]  -  29\n",
    "\n",
    "\n",
    "\n",
    "* C5 \n",
    "    * C5 - d1 (& d2): PTMA, *ALX4*, *CRABP1*, *SDC1*, *SPON1*, *TNN*, ***BMP7***, ***FBXO32***, ***IGFBP3***, ***INHBA***, ***MRPS6***, ***PGM2L1***, ***RSPO3***, ***TFAP2A***, ***TRPS1***  -  [1, **0**, *5*, ***9***]  -  15\n",
    "\n",
    "\n",
    "* C2 and C3 vs c/d: C2 might be more clear than C3 to be associated to c/d, possibly because C3 is also a bridge cluster. \n",
    "    * C2 - c/d: NR2F1, PTH1R, SRPX, **CPNE5**, **CRABP1**, **MEOX2**, **NCAM1**, **RSPO4**, *CHST15*, *CYP1B1*, *TBXA2R*, *TRIB2*, ***CCK***, ***COCH***, ***FIBIN***, ***FMOD***, ***MKX***, ***PLXDC1***, ***PTGFR***, ***TNMD***  -  [3, **5**, *4*, ***8***]  -  20\n",
    "    * C3 - c/d: TPM2, **COL8A2**, **MFAP2**, **TRIL**, *EGFLAM*, *NREP*, *RFLNB*, ***COL7A1***, ***F2R***, ***MMP16***, ***RASL11B***  -  [1, **3**, *3*, ***4***]  -  11\n",
    "    * C2 - C3 - c/d: COL11A1, KIF26B, NRP2, TRPS1, **EMID1**, **TBX15**, *ADAMTS9*, *EDNRA*, *TSHZ3*, ***DKK2***, ***GPM6B***, ***MAFB***, ***TCF4***, ***TENM3***  -  [4, **2**, *3*, ***5***]  -  14\n",
    "\n",
    "---\n",
    "\n",
    "It looks like b5 and e1 are related to either D1 and D2, but b5 is more related to D1, while e1 to D2\n",
    "\n",
    "* D1 and D2 vs b5\n",
    "    * D1 - b5: CNN3, LUM, **ABCA8**, **ETV1**, **PHLDA1**, *ITM2A*, *SPARCL1*, ***APOD***, ***COL8A1***, ***ENTPD2***, ***GPC3***, ***P2RY14***, ***SOX9***  -  [2, **3**, *2*, ***6***]  -  13\n",
    "    * D2 - b5: BHLHE40, PHLDA3, **CSRP1**, *DDIT4*, ***SBSPON***,  -  [2, **1**, *1*, ***1***]  -  5\n",
    "    * D1 - D2 - b5: KLF5, **CCL2**, **MEOX2**, **TGFBI**, **TM4SF1**, *MATN2*, ***CLDN1***, ***EBF2***, ***FOXS1***, ***NR2F2***, ***VIT***  -  [1, **4**, *1*, ***5***]  -  11\n",
    "    \n",
    "     \n",
    "* D1 and D2 vs e1  [En e1 y b1 también salen bastantes genes de d2]\n",
    "    * D1 - e1: CTNNAL1, PTCH1, *EGR3*, *HMGA1*, ***ETV4***, ***SOX9***, ***TIAM1***  -  [2, **0**, *2*, ***3***]  -  7\n",
    "    * D2 - e1: CCDC3, MTSS1, **CSRP1**, **LMO7**, **SBSPON**, *CAVIN2*, *FRMD4B*, *PHLDA3*, *SYNE2*, *TPD52*, ***BNC2***, ***CAV1***, ***CAV2***, ***EFNB1***, ***ITGA6***, ***ITGB4***, ***KRT19***, ***PALMD***, ***SLC2A1***  -  [2, **3**, *5*, ***9***]  -  19\n",
    "    * D1 - D2 - e1: EZR, **NR2F2**, *KLF5*, ***CLDN1***, ***EBF2***, ***NDRG2***, ***TENM2***   -  [1, **1**, *1*, ***4***]  -  7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.system('jupyter nbconvert --to html 7HM_mouse_human_comparison.ipynb')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:triku-notebooks]",
   "language": "python",
   "name": "conda-env-triku-notebooks-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "384px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}