help with germline analysis using cnvkit

Question

[ww22runner]

Hi all, my bio is not very strong so please excuse this basic questions. I am trying to process only germline samples in my run using cnvkit. Usually with paired samples, I run:

cnvkit.py batch *Tumor.bam --normal *Normal.bam \
    --targets my_baits.bed  \
    --fasta hg19.fasta --access data/access-5kb-mappable.hg19.bed \
    --output-reference my_reference.cnn --output-dir results/ \
    --diagram --scatter

with just germline samples I should run:

cnvkit.py batch --normal *Normal.bam \
    --targets my_baits.bed  \
    --fasta hg19.fasta --access data/access-5kb-mappable.hg19.bed \
    --output-reference my_reference.cnn --output-dir results/ \
    --diagram --scatter

and

cnvkit.py batch $CNV_BAMS/*_G.bam \       
 -r  $CNV_BAMS/my_reference.cnn  \ 
 -p 8 \
-d $CNV_G_RESULTS

right?

Also, cnvkit Version v0.9.7 beta suggests that after the batch step, we should call these commands: a) segmetrics b) call c) bintest d) cnv_ztest.py

Could someone help me understand in simple terms what each step post batch command means and what the outcome is for that step?

Thank you

Answer 1

[Eric T.]

The steps of the analysis are generally:

1. Create a reference profile of the sequencing coverage variation that occurs in samples that don't have large-scale or recurrent copy number alterations. This step just creates the reference profile, not a set of copy number calls.
2. For "tumor" or "test" samples of interest, where you do want find copy number alterations, subtract the reference profile (etc., some statistics) to remove the variation that's likely due to sequencing coverage rather than real genomic alterations. This creates the ".cnr" file.
3. Analyze the .cnr log2-ratio values to find a reasonably parsimonious set of piecewise-constant segments. This creates the ".cns" file.
4. Reanalyze the .cns segments to remove likely artifacts and distinguish copy-neutral segments from likely real genomic alterations.