tetedange13
commented
2 years ago Hi @xinmiaoyan ,
Not an author of CNVkit, but I may help => Several points:
- I saw looking at your other issue that you are running CNVkit v0.9.6 --> Please first update because that version is pretty old and lot of things changed / were fixed since
- Based on CNVkit documentation: in a paired tumor-normal situation CNVkit gives better results if you pool all your normal samples into a single reference then call CNV on each of your tumor ones => If not already done, you should carefully read CNVkit's very complete documentation
- Looking at your BAM name, I guess you removed duplicates reads ? Sometimes this can have a negative impact on results, so if possible try also without removing duplicates (after doing previous things I mentionned)
=> See this answer from one of CNVkit collaborators
Hope this helps ! Kind regards, Felix
hello, I ran cnvkit with paired tumor-normal bam files, and get weird result. among the 40 paired samples, every one is looks like this, the left bar is always blue. is there something wrong? Do we need to build the reference.cnn by all the normal samples or just the paired normal sample? Thanks for your response.
here's the code:
cnvkit.py batch tumor_485700.sorted.rmdup.bam\ --normal normal_485700.sorted.rmdup.bam \ --targets Regions_nohead.bed \ --fasta hg38.fa\ --access access.hg38.bed \ --annotate refFlat.txt \ --output-reference my_reference.cnn \ --output-dir result \ --diagram --scatter