search

jodyphelan / TBProfiler

Profiling tool for Mycobacterium tuberculosis to detect ressistance and strain type from WGS data
GNU General Public License v3.0
105 stars 43 forks source link

QC metrics #377

Closed thanadon1437 closed 4 months ago

thanadon1437 commented 5 months ago

Hi @jodyphelan,

I profiled my tNGS data obtained from nanopore sequencing using TB-profiler and it presented two variants in the rpoB gene: L430P (761,095) and D435A (761,110) both variants associated with resistance. One variant was reported in the “resistance variants report,” while the other was listed in the “QC failed variants report.”

Additionally, there are a few samples where resistance-associated variants were found, which are concordant with the pDST, but these variants were also reported in the QC failed section.

Could you please explain the criteria or QC metrics that TB-profiler uses to determine which variants pass or fail quality control?

Thank you!

jodyphelan commented 5 months ago

Hi @thanadon1437 Usually this means that it is too low in depth. Have a look here for info on the variant filtering: https://jodyphelan.github.io/tb-profiler-docs/en/algorithm/#variant-filtering Additionally see here for more info: https://github.com/jodyphelan/TBProfiler/issues/239#issuecomment-2054068666

thanadon1437 commented 5 months ago

Hi @thanadon1437 Have a look here for info on the variant filtering: https://jodyphelan.github.io/tb-profiler-docs/en/algorithm/#variant-filtering Additionally see here for more info: #239 (comment)

What could be the reason of soft-fail in my case since I am using single-end data, having over 95% estimated fraction, and those variants are also located very close to each other?

Would you mind to guide me calculate the depth in that region? Can I use samtools depth?

thanadon1437 commented 5 months ago

I just ran samtools depth -r Chromosome:760000-761300 input.bam > depth_rpob.txt

the one that was soft fail shown Chromosome 761095 99

and another that reported Chromosome 761110 101

I attached the output of samtools depth here depth_rpob.txt

jodyphelan commented 5 months ago

Could you post the json result file?

thanadon1437 commented 5 months ago

Could you post the json result file?

Sure, I am willing to know more about the explanation. here the .json file barcode01.results.json

and this also the sample that 2 rpoB variants and 1 inhA (both are WHO G1) reported in QC failed barcode05.results.json

Thank you