SNP array CNV calling for Cancer tissue (.CEL files)

Question

[German.M.Demidov]

Hi there,

I was working for 5 years with NGS data and CNV detection, but now I have to perform CNV calling in .CEL files from Affymetrix v6 arrays. At first, I have no idea how to operate with such files (should I download some converter to vcf or txt or other format?). At second, I need to perform tumor CNA calling there (of course I have tumor-normal data). Which tool is the best, taking into account purity and clonality of samples?

Thanks a lot for any hints!

(of course I googled for papers, but I am a complete newbie in this field and a bit overwhelmed, everybody says 'oh we are the best buy our software', but somehow I dont trust most of them and also I work in academia and looking for something which is free for research)

[Kevin Blighe]

Related answer: A: What is the best pipeline to analyze affymetrix snp 6.0 array in 2020?

Answer 1

[Kevin Blighe]

Around the time when NGS was becoming very popular, the then Affymetrix CEO, Frank Witney (I managed to attend a meeting with him once), still firmly wanted to focus the company on microarrays. At the time, they [Affymetrix] had been pushing their OncoScan product. Prior to this was the Affymetrix SNP 6.0 array (SNP 6.0 array), which has probes for both SNPs and CNVs, as you know. The leading method that emerged for calling copy number segments in this SNP 6.0 array was Circular Binary Segmentation (CBS). The Aroma project did a lot of work in producing materials and guides for, for example, calling somatic copy number segments between tumour and normal. Take a look: https://aroma-project.org/vignettes/

Their code should still work. You will require the CDF for the SNP 6.0 Array, though. If you do not have it, it should be available on ThermoFisher's website (they bought Affymetrix). I may also have it on my disk somewhere.

Kevin

[German.M.Demidov]

Thanks a lot! I will try Aroma! What is CDF? For me it rings a bell only about cumulative distribution function...

[Kevin Blighe]

Hey, no, in this context, it means Chip Definition File. It contains information about the layout and design of the probes, the raw fluorescent values for which are stored in the CEL files. There is more here: http://aroma-project.org/definitions/chipTypesAndCDFs/

Answer 2

[Giulio Genovese]

If you are interested in clonality, you can use the affy2vcf bcftools plugin to convert the apt-probeset-genotype output to VCF and then use the MoChA pipeline to call mosaic chromosomal alterations arising from clones