Skip to content

Latest commit

 

History

History
48 lines (38 loc) · 3.76 KB

README.md

File metadata and controls

48 lines (38 loc) · 3.76 KB
Raw

WisecondorXtena

A modified version of the copy number caller, WisecondorX (https://github.com/CenterForMedicalGeneticsGhent/WisecondorX), that now infers purity and produces improved copy number calls from lpWGS data.

The modified/improved version of WisecondorX

This version of WisecondorX utilizes the original code with no specifications other than an elevated bin size (the default is 5kb, this uses 500kb), calculates purity and how offset the mean log2 ratio is from zero, and reiterates these values back into the modified code as parameters that adjusts the genomic segments accordingly.

The original WisecondorX consists of 3 main steps

  • File conversion (aligned reads (BAM) to NPZ file format)
  • Reference creation (using reference NPZ files)
  • CNV prediction (through a user-definable cut-off ("beta"), or the default (leverages Z-scores))

WisecondorXtena is run via snakefile, consisting of two sections

The run_updated_wisecondorX.smk set up:

Section 1 (initial run to calculate offset mean and purity):

  • Runs the 3 WisecondorX steps with the original parameters and specifies 500kb bins.
  • Plots the copy number profile that would be originally output by WisecondorX prior to any modifications made (this is called "unshifted") along with the corresponding density plot from a Gaussian mixture model.
  • Determines how offset the mean log2 ratio is from zero (corresponding to how biased the assigned genomic aberrations are from the copy neutral state), as well as estimates the purity of each sample.

Section 2 (second run of WisecondorX to adjust segments):

  • Reruns WisecondorX the offset value and purity estimate and inputs those as the offset_neut_peak and beta paramters, respectively.
  • Produces new copy number profile plots with the density curves, after the adjustment is performed.
  • A new purity estimate is calculated. This can be compared to IchorCNAs tumour fraction value.

To note:

  • The offset line in the density curve and the offset mean value are both no longer relevant after shifting is complete and the second set of plots are generated.
  • Aberrant segments that are amplified are displayed only when the argument beta is used within Section 2 of the WisecondorX code. When beta is not used, Z-scores are utilized to call copy number aberrations and therefore only lost, neutral, and gained segments are assigned.

New optional argument

(See the WisecondorX github page for the pre-existing paramaters and their description)

--offset_neut_peak - A value calculated by the mixture model (using 1 cluster) that can be positive or negative (positive will cause an upwards shift and negative will cause a downwards shift in genomic segments).

  • How it works : this value is the mean log2 ratio of the neutral peak from the mixture model. It is added to the results_r, results_z and results_w variables to shift resulting log2 ratios and plots.

New file generated

  • A segment file (.seg) is now produced by WisecondorXtena

WisecondorXtena directory structure for snakefile

This can be modified to the users preference, but this is the structure that Kristena has.

  • wisecondorX (main folder)
    • lpWGS_all_500kb_bincorrection
      • 00_reference_bam (contains bams to be used for reference creation)
      • 01_reference_npz (contains reference bams converted to npz format)
      • 02_samples_npz (contains sample bams converted to npz format)
      • 03_modified_code_output_all
        • 01_unadjusted_output (results from the first run of wisecondorX, before segments are adjusted)
        • 02_adjusted_output (results from the second run of wisecondorX, after segments are adjusted)
    • lpWGS_MarkDuplicates (contains samples to be run through WisecondorXtena, these have gone through Chris' ctDNA pipeline including the Picard Mark Duplicates program)