sciClone input vaf file?

Question

[Haitao]

Dear All,

Hi, I want to use sciclone on our exome sequencing data. but one thing I can't understand that is how can I got varCount equal to 0? I have no idea about this,

following data i just grep from sciclone-meta-master manuscript figure3 data vaf input:

refCount varCount
261 0
102 0
1462    0

so if that is the case, I totally wrong? I used GATK merged vcf, and use R (data.table) to manipulate and got final input files.

I just wonder if anyone can help me and give me some advice. and would you mind to teach me how to calulate (vcf ./.) the depths for those paired samples only one sample have mutaiton but other one no mutation? bedtools coverage? or other ways?

Thanks so much, Haitao

Answer 1

[Chris Miller]

Often a mutation will only be present in one sample and not the other. This is not unexpected. We typically use bam-readcount to pull VAFs from both bams for all sites.

[Haitao]

Dear Miller,

Thanks so much for you reply so quickly, I will try this bam-readcount to pull VAFs from both bam.

One more question: for your CNV segment file, I found those distance of segment windows are so diverse in different chromosome:

chr1 1 351000
chr1 351001 650000
chr2 1 215000
chr3 1 500000
...

I just wonder which tools you used? Actually, I use VarScan, Conifer and CNVkit to call CNV, segmented by DNAcopy. but when I got segment CNV should I convert by Bedtools -w 50k (or other num) to make window distance, and re-calulated again? or just used DNAcopy segment CNV as input file for sciclone? before input to sciclone should I remove those CNV changed region or I can give all the info to sciclone, later use copyNumberMargins=0.25 to filter out?

Also I saw you CNV segment value already changed back to real CNV copy number: like (2^0.5)*2, but in R script still use

 copyNumberMargins=0.25

Thanks for you help and your tools are really really nice and help our team a lots. Haitao

[Chris Miller]

• Yes, it's real copy number (2 = cn neutral, 3 - amp, etc). 0.25 is the distance away from 2 that is allowed for CN call to be considered neutral
• The results are from varscan exome, run through the DNAcopy package for segmentation. It identifies regions of concordant CN, so yes, the lengths will be variable. Any similar CN and segmentation algorithm should be fine.

[Haitao]

I will modify my pipeline based on your suggestion. Thanks! Haitao

[Haitao]

Dear Chris,

I'm following what you mentioned in this post, I got bam-readcount output file now:

GL456392.1      772     C       123     =:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  A:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  C:103:30.22:36.83:1.46:46:57:0.57:0.06:145.63:46:0.46:64.46:0.33        G:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  T:20:31.10:37.00:0.00:11:9:0.78:0.09:182.55:11:0.59:55.90:0.34  N:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
GL456378.1      780     T       2       =:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  A:2:33.00:37.00:0.00:1:1:0.32:0.04:129.50:1:0.71:100.00:0.38    C:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  G:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  T:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00  N:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00

However, I don't know which tool can help me to convert bam-readcount output to like this VAF output:

chrom POS REF ALT REF_DP ALT_DP TOTAL_DP
GL456378.1 772 C A 103 20 123
GL456378.1 780 T A 0 2 2

Thanks in advance, Haitao

Answer 2

[Haitao]

Hi Chris,

I have one more question for building clonal evolution, since when I use sciclone to calulate subclone, I already remove those un-neutral region, after got subclone info, can I add those CNV changed region back to build evolution event? I ran THetA2 succesfully, I plan use sciclone + THetA2 to test the evolution. plot by fishplot. Thanks. Haitao

[Chris Miller]

There are options like Theta that allow for merging that information back in, but no supported procedure for merging the two that I know of. You're welcome to invent one, as it's probably possible.

Answer 3

[2891913164]

Hi Chris, I also want to know how to convert the bam-readcount output file to the vaf data required by sciclone. I hope to get your help! thank you very much!

[Chris Miller]

One way is to use something like VATools to annotate your VCF with that information, then extract it by parsing the VCF. https://github.com/griffithlab/VAtools