Question: Visualizing Cancer Heterogeneity Results From Ngs Data
1
gravatar for newDNASeqer
3.0 years ago by
newDNASeqer520
United States
newDNASeqer520 wrote:

I am working on cancer heterogeneity NGS data derived from patients (primary) solid tumors, and generations of mice that carry the cancer patient's primary tumors. I did exome-seq analysis and have identified SNPs/Indels on many genes in the cohort.

The passage from primary patient's sample to the third generation (F3) of mice shows that the tumor accumulates more mutations, but the mutational patterns are also heterogeneous among the same generation of mice. I am wondering about an effective way for presenting the heterogeneity results. Some methods I was thinking about include:

  1. Get the ratio of the variant allele frequency (VAF, each mutation identified) in each mice vs the primary tumor (patient), then calculate the correction coefficient between the mice and the primary sample; plot the VAF vs sample lineages (?)

  2. Sum up the number of SNPS/Indels in each sample, and divide the number by the total number of mutations. This will give a mutation rate in each sample, then plot this numbers against each sample name. I would expect this to show a linear accumulation.

I am not sure these two presentations will make any sense, so I am looking for suggestions for presenting the heterogeneity results. If you know some methods or have seen some good way of presentation, please reply. I appreciate it.

ngs visualization cancer • 1.5k views
ADD COMMENTlink modified 3.0 years ago by Chris Miller17k • written 3.0 years ago by newDNASeqer520

hi, can you please clarify some sentences of your question:

"The passage from primary patient's sample to the third generation (F3) of mice shows that the tumor accumulates more mutations" -> more mutations compared to what? to the original tumor, or to the number of mutations you would expect in a third generation?

", but the mutational patterns are also heterogeneous among the same generation of mice." -> what does this mean, exactly?

1) typically, the mutational patterns within the same generation of mice are heterogeneous

2) The mice obtained from the same patient accumulate different spectra of mutations

3) different replicas of the same experiment show different patterns

4) something else

Moreover, it would be easier to suggest you a good visualization if you can clarify better what the null hypothesis of your analysis is.

ADD REPLYlink modified 3.0 years ago • written 3.0 years ago by Giovanni M Dall'Olio25k

Sorry for being not clear on that. To clarify:

The cancer patient's primary solid tumor (called F0) was chopped into multiple pieces, and each piece is transplanted to a mouse (called F1). Let the tumor grow in the mice, and when the tumor grows to a certain size, we sacrifice the tumor and get the tumor. The tumor is then chopped, and each piece is transplanted to the next generation of mice (called F2). Repeat this process until we get the third generation of mice (called F3) that carry the solid tumor.

With the tumor samples derived from patient, and the three generations of mice, we performed exome-seq. The bioinformatics part involves detecting the SNPs/Indels, and copy number changes.

The genetic variants results indicate that some genetic mutations (e.g. BRAF V600E) are found only in a subset of the mice of the same generation, while other mice may have different mutations or WT on the same gene (e.g. BRAF Wild-type)

My question is how to present the genetic heterogeneity results effectively.

Thank you

ADD REPLYlink written 3.0 years ago by newDNASeqer520

This is an interesting question and not a straightforward one. I still think plotting the vafs of individual variants in the tumor against the same from the passages would be helpful. Here are two previous studies on similar topics from our group that may give you some inspiration: http://www.cell.com/cancer-cell/fulltext/S1535-6108(14)00054-3 and http://www.cell.com/cell-reports/fulltext/S2211-1247(13)00463-4

ADD REPLYlink written 3.0 years ago by Chris Miller17k
1
gravatar for Chris Miller
3.0 years ago by
Chris Miller17k
Washington University in St. Louis, MO
Chris Miller17k wrote:

I'm not entirely clear on what question you're trying to answer, but the SciClone R package may be of interest to you, (specifically the 2d or 3d plots).

ADD COMMENTlink written 3.0 years ago by Chris Miller17k

Thanks Chris. You are the author of SciClone? great work! Good luck to your submission on PLOS Computation.

One quick question for you on calculating VAF (Variant Allele Frequency): how did you calculate the VAF in your plotting (shown on github)?

Thanks

ADD REPLYlink modified 3.0 years ago • written 3.0 years ago by newDNASeqer520

One straightforward way is to take the variants output by your pipeline and run them through bam-readcount. The VAF is the number of variant supporting reads divided by the total number of reads.

ADD REPLYlink written 3.0 years ago by Chris Miller17k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2008 users visited in the last hour