gistic for whole genome sequencing copy number analysis

Question

[owen]

Hello, I am analyzing whole genome sequencing data from cancer samples.

For copy number analysis, I have two questions:

1. How can I convert chromosome segments to gene level copy numbers?
2. How can I apply GISTIC to segmented copy number data? I want to identify recurrent copy number changes in cancer population, but GISTIC is originally desinged for array data with probes.

Answer 1

[Irsan]

1) The simplest way is to use the CNtools bioconductor package. Other more flexible but more intensive option is to use genomic overlap tools like bedtools, bedops or GenomicRanges to map back the segments to your gene/transcript annotation of interest

2) If you have segmented copy numbers, a file describing the genomic windows that were used to quantify the reads per sample and a reference genome than you have everything thats required as input for GISTIC2. Also see GISTIC2 documentation

[owen]

1. Thanks for your help. Results of CNtools is genomic segment, but not gene-centric. Other tools seem to work.
2. How can I find the file for genomic windows? Is it in the _ratio.txt file?

[rse]

Hi owen, were you able to run gistic?

[rse]

Hi irsan, i have the segmented file. But how to get the number of markers?

[Gonzalo]

I ran into the same problem. My solution was to create a marker file which included markers at every segment boundary plus markers every 2Kb across the whole genome. Did you solved the problem differently?

[benche]

Hi Irsan, I am trying to run the GISTIC2.0 in the Genepattern to analyse the segmented results generated by the CBS.However, I always get an error "Index exceeds matrix dimensions", and the possible measures given by the website says "check your markers file format. Please see the sections on the markers file format in the GISTIC documentation for more details and examples." I carefully examine and compare the format of my datasets with the samples' and I fail to find any difference between them.So I urgently hope that you can do me a favor and give me some advices in your convenient time. The segmentation file, output of the CBS, removes the first title row and the first serial number column, and saves in the tab-delimited txt file. Then the marker file made by myself holds three columns ,SNP marker label, Chromosome number and Chromosome position, that is a subset of marker annotation file of SNP array before segmentation and saves in the same mathod with segmentation file. It's my data processing measure, and I desire to get your instruction or teach me how to get the input files of the GISTIC.

[Dmitri Ivanovich]

Hi Benche, It seems that I have met similar problem. Have you solve yours?

[sutturka]

I have called somatic CNVs using GATK4 on canine (dog) data. I have standard output from GATK which is a seg file as shown below

I am not sure how should I convert above format to GISTIC input format. Any comments would be helpful.

Also, for the markers file this link suggested a simple format for markers file:
sample_name chr start_pos
sample_name chr stop_pos

Do you think would be the correct marker file to use.

[sishuinianhua0502]

Hi Sutturka,

I also called somatic CNVs using GATK4. It seems that I have met similar problem. Have you solve yours? I urgently hope that you can give me some advices in your convenient time if you have completed it. Thank you very much.