It may (apparently) happen that Kallisto with the customized index for a given patient will not count any transcripts mapping to any allele of a gene (i.e., the total transcript count for a gene is zero), which will lead to errors during baf computation due to division by zero (see below).
|
baf = allele_results[gene]["allele1_count"] / ( |
|
allele_results[gene]["allele1_count"] |
|
+ allele_results[gene]["allele2_count"] |
|
) |
Easiest mitigation would be to just check and report NAs for that gene (probably).
It may (apparently) happen that Kallisto with the customized index for a given patient will not count any transcripts mapping to any allele of a gene (i.e., the total transcript count for a gene is zero), which will lead to errors during baf computation due to division by zero (see below).
arcasHLA/scripts/quant.py
Lines 282 to 285 in 8ba97fb
Easiest mitigation would be to just check and report NAs for that gene (probably).