tlesluyes
commented
2 years ago Hi @shmuen,
You're right and that is the expected behaviour of the ascat.prepareHTS function: normal logR is just set to 0 for all probes. Coverage in the normal is used to compute tumour logR (and that is very important) but then normal coverage/logR is not considered at all so we artificially set it to 0 (and it's not used in the code). This is because ASCAT leverages genomic information (logR/BAF) at heterozygous SNPs (in the normal) to resolve gains, losses and CN-LOH (based on the logR/BAF agreement). Therefore, BAF in the normal is essential and will be used whereas logR is propagated for plotting (so our functions work for both sequencing data and microarrays).
Of note, we're about to update our methodology to process targeted sequencing data because we can make use of the available allele counts to genotype SNPs based on a probabilistic method rather than hardcoded thresholds. Also, cohorts with a few females (typically less than 10, don't know whether this applies to you) will have very few SNPs on the nonPAR region on chrX so we are in the process of rescuing a few if the data is clean enough. This should be available in a few weeks from now.
Cheers,
Tom.
shmuen
Hey,
I am using ASCAT to call copy numbers and purity from targeted sequenzing data (using the workflow described here: https://github.com/VanLoo-lab/ascat/tree/master/ExampleData#processing-targeted-sequencing-data). When executing prepareHTS , the resulting normalLogR_file contains only zeros. After checking the code, I realized that normalLogR is initialized (https://github.com/VanLoo-lab/ascat/blob/bf325ddff3c1f3af93e05bbea950ec197904368d/ASCAT/R/ascat.prepareHTS.R#L126), but never used after (except turning it in a data.frame) and only written as table in line 162. Is this intentional or am I missing something?