Infer CNVs from SNP variant allele fractions

Question

[ceruleanivy]

I am currently working on data obtained from proton sequencing (IonTorrent) on bulk tumor micro FFPE samples, using panels for exome targeting. Although I've came across some literature on that issue I am not quite experienced so I would like to ask what are the most popular algorithms/software on inferring copy number variations (CNVs) from single nucleotide polymorphism (SNP) variant allele frequencies (VAFs). Here's an example of my data

Answer 1

[willgilks]

Hi ceruleanivy,

In R you might be able to do this using dplyr/tidyr functions.

1. Convert the data to long format (sample-snpid-depth1-depth2-genotype).
2. Calculate the allelic depth ratios in a new column.
3. Standardise the distributions of the two alleles.
4. Split data fram by genotype
5. group_by(snpid)
6. Test for standardised depth ratios for normal distribution.
7. Bin the results into windows of 100Kb first, to see if any big deletions/duplications show-up.

..or try this program https://sourceforge.net/projects/excavator2tool/ published a few months ago https://academic.oup.com/nar/article/44/20/e154/2607979/Enhanced-copy-number-variants-detection-from-whole