I have low-coverage WGS data from embryonic tissue. The exact number of cells is unknown (4-6 cells taken from early embryonic tissue) . Due to the possibility of embryonic mosaicism, the sample could have germline mutations, somatic mutations, both, or neither. I am trying to develop a pipeline to determine what fraction of the sample, if any, has some kind of aneuploidy.
What would be the best approach to doing this? I originally considered building some model that models the peaks in BAF as a mixture of clonal BAF patterns, but I would need quite a bit of work to train, tune, and validate a bespoke model. I'm wondering if there is a good out-of-the-box solution.
We already use CNVkit in our pipeline and so I'm considering using the .cns file from sample analysis (with big bins to account for the low coverage), taking the log2 values stored therein, and passing them to PureCN. However, I'm not sure if my situation (e.g. fairly simple chromosomal aberrations, the possibility that all cells are aneuploid) makes PureCN a good fit.
I wanted to know if this is an appropriate approach to take, and if not, what might be better.
Thank you!
