search

parklab / NGSCheckMate

Software program for checking sample matching for NGS data
MIT License
125 stars 48 forks source link

count_alt_A_C_G_T #52

Closed itaibeno closed 4 months ago

SooLee commented 4 months ago

Thanks for the PR but I wouldn't merge this one since it changes the behavior different from the original algorithm.

itaibeno commented 4 months ago

@SooLee thank you for reviewing it. the main goal is to count the alt alleles in order to reuse the tool when the alt alleles were changed enzymatically and NOT due to mutation so one can apply correction. printing the distribution of the total alt reads can be helpful.

In addition one issue I found when analyzing the pattern file, there are duplication of Left right center 21mers between different loci. e.g. index 510 and 3405 hold the same 21mer so the code actually count the 21mer to 3405 the latest index in hash.

in that case the vafs are not correct for 510. it is small number 215 out of 11696 but it is not correct.

itaibeno commented 4 months ago

duplicates_loci.251.info.csv