Skip to content

Latest commit

 

History

History
51 lines (33 loc) · 3.12 KB

README.md

File metadata and controls

51 lines (33 loc) · 3.12 KB
Raw

gene-tools

Tools to reliably map protein IDs to gene names, made during my UROP at the Lage Lab at the Broad Institute of Harvard and MIT.

map

The main project - automating the assignment of protein IDs / accession numbers to HGNC gene names. Takes in a list of IDs as input, separated by newlines, and returns a list of assigned gene names where possible, and provides information about all other cases. For instance, map reports unassigned protein IDs and returns its Ensembl ID wherever possible.

Setup

The tools draw information from UniProt and HGNC, both locally and programmatically through queries. To set up the local databases properly, follow the instructions in the following folders: ./human_data/ and ./hgnc_data. These are repeated below for completeness.

  1. ./human_data: From the UniProt Downloads page, go to 'Taxonomic Divisions' and download the uniprot_sprot_human.dat.gz and uniprot_trembl_human.dat.gz files. Extract them into the ./human_data folder. Then run grep.sh to create the data.txt file.

  2. ./hgnc_data: From the HGNC Custom Downloads page, download two files:

  3. Download a file with only the 'Approved Symbol' and 'UniProt ID' checked. Save this as hgnc_symbol_ac.txt in the ./hgnc_data folder.

  4. Download another file with only the 'Approved Symbol, 'Previous Symbols' and 'Synonyms' checked. Save this as hgnc_symbol_previous_synonym.txt in the ./hgnc_data folder.

At the end of this setup, your directory should look as so (the map and match directories are not expanded, they should not be modified):

gene-tools
- hgnc_data
--- hgnc_symbol_ac.txt
--- hgnc_symbol_previous_synonym.txt
- human_data
--- data.txt
--- grep.sh
--- uniprot_sprot_human.dat
--- uniprot_trembl_human.dat
- map
- README.md

Usage

Note that using map requires an Internet connection, since some queries will be resolved online.

map

map takes input from ./map/in.txt, which is a list of UniProt IDs or Accession Numbers, and outputs ./map/results.txt, a tab-spaced list of those IDs with corresponding HGNC gene names, along with status flags that indicate how the gene name was obtained. For instance, the ID Q15465 will be mapped to SHH directly on HGNC.

map also identifies problematic cases (i.e. cannot be mapped solely on HGNC) and resolves them accordingly where possible. For instance,

  1. Obsolete IDs: IDs that were once in use, but not any more. e.g. E9PEB9 is an obsolete ID on UniProt and cannot be found on HGNC - map will tell you that the last existing gene name on UniProt is DST and checks if DST is the correct HGNC gene name (it is).

  2. Unassigned IDs: IDs that exist, but have not been assigned a gene symbol. e.g. P00761 does not have an assigned gene name - map reports this, and returns its Ensembl ID where possible.

  3. Bad IDs: IDs that do not exist, possibly as a result of a typo.

  4. Not found in HGNC: IDs that can be mapped in UniProt but not on HGNC. map reports the UniProt gene name instead.