kircherlab / cfdna_gccorrection Goto Github PK

In bam files with low coverage (e.g. 1X), it might happen that no fragments were observed for a given fragment length.

These cases need to be handled for calculating the Bias ratios and scaling parameters.

When reading parameters from precomputed background file, make sure that the seed is set for all random number generators otherwise reproducibility issued might arise

There should be a check for precomputed background files, if no blacklist was provided.

This includes that ther was no blacklist specified in both cases!