broadinstitute / gatk

Official code repository for GATK versions 4 and up
https://software.broadinstitute.org/gatk
Other
1.71k stars 590 forks source link

PathSeq Host Alignment Parameters #6819

Open wir963 opened 4 years ago

wir963 commented 4 years ago

Feature request

Tool(s) or class(es) involved

PathSeq v4.1.8.1

Description

There are two possible alignment metrics for the microbial alignment --min-score-identity and --bwa-score-threshold. However, there is only one possible alignment metric for the host alignment --host-min-identity. If a user wants to use --min-score-identity, which is a function of the read length, there is no function of the read length metric for the host identity to ensure that (at a minimum) the read aligns better to the microbial genome than the host, which is very important.

mwalker174 commented 4 years ago

Thanks for your suggestion @wir963. The --min-score-identity and --host-min-identity parameters can be used to tune your desired specificity/sensitivity for microbe read detection. The default settings should guarantee that the identified microbial reads have better alignments to the microbe reference than host.