Empty from sRNAbench output (genome mapped output)

kicheolkim commented 4 years ago

Expected behavior and actual behavior.

I tried input using output from sRNAbench genome mapping mode. I expected gff output with miRNA/isomiR information and counts, but the file was empty like this:

## mirGFF3. VERSION 1.2
## source-ontology: miRBasev22 doi:10.25504/fairsharing.hmgte8
## TOOLS: s,r,n,a,b,e,n,c,h
## COLDATA: S400001603

Here is the log file:

04/16/2020 09:17:37 INFO-mirtop.libs.logger(29): Run annotation
04/16/2020 09:17:37 INFO-mirtop.libs.logger(48): Reads with isomiR information 7804
04/16/2020 09:17:37 INFO-mirtop.libs.logger(142): Loaded 0 reads with 3040 hits
04/16/2020 09:17:37 INFO-mirtop.libs.logger(143): Reads without precursor information: 11124
04/16/2020 09:17:37 INFO-mirtop.libs.logger(145): Reads with MV as variant definition, not supported by GFF: 1101
04/16/2020 09:17:37 INFO-mirtop.libs.logger(146): Hit Filtered by having > 3 changes: 6703
04/16/2020 09:17:37 INFO-mirtop.libs.logger(48): Reads with isomiR information 5254
04/16/2020 09:17:37 INFO-mirtop.libs.logger(142): Loaded 0 reads with 1354 hits
04/16/2020 09:17:37 INFO-mirtop.libs.logger(143): Reads without precursor information: 6907
04/16/2020 09:17:37 INFO-mirtop.libs.logger(145): Reads with MV as variant definition, not supported by GFF: 600
04/16/2020 09:17:37 INFO-mirtop.libs.logger(146): Hit Filtered by having > 3 changes: 4654
04/16/2020 09:17:37 INFO-mirtop.libs.logger(48): Reads with isomiR information 6618
04/16/2020 09:17:38 INFO-mirtop.libs.logger(142): Loaded 0 reads with 2493 hits
04/16/2020 09:17:38 INFO-mirtop.libs.logger(143): Reads without precursor information: 8730
04/16/2020 09:17:38 INFO-mirtop.libs.logger(145): Reads with MV as variant definition, not supported by GFF: 804
04/16/2020 09:17:38 INFO-mirtop.libs.logger(146): Hit Filtered by having > 3 changes: 5814
04/16/2020 09:17:38 INFO-mirtop.libs.logger(57): It took 0.010 minutes

Steps to reproduce the problem.

It's working with the output from library mapping mode of sRNAbench with same sample. I saw the contents in gff output. But the gff was empty only when I used genome map mode output.

Specifications like the version of the project, operating system, or hardware.

I'm using mirtop 0.4.23, and sRNAbench version is 2.0 - 4/2020

Thanks!!

lpantano commented 4 years ago

Hi, Thanks for the feedback, I think this line is telling something interesting:

04/16/2020 09:17:37 INFO-mirtop.libs.logger(143): Reads without precursor information: 11124

Can you send the command used?

As well, can you check whether the miRNA in the sRNAbench matches the names in the GFF from miRBase you are using?

Just to be sure that is not a bug in master:

Can you try the devel version of mirtop?

Thanks!

kicheolkim commented 4 years ago

Thank you for the quick reply!! I've run with develop version but it's still empty.

I used fasta and gff3 files downloaded from the miRBase website, so, I believe they match. And it works well with output from a different mode. (add: confirmed name match between hsa.gff3 and hairpin.fa)
I used command as below:

mirtop gff --sps hsa --hairpin /Users/kicheol/working/pipeline_dev/smallRNA/Test_Qiaseq/output/test_v2.2_isomiR/hairpin.fa --gtf /Users/kicheol/working/pipeline_dev/smallRNA/Test_Qiaseq/output/test_v2.2_isomiR/hsa.gff3 --format srnabench --out-format gff -o ./ ./S400001603 ./S400001616 ./S400001617
And here is the log file with '-d' option in case it's helpful for you.

lpantano commented 4 years ago

Hi, thank you for the log file, for some reason the files has Chr information there as well. So mirtop is seeing this line:

mature#hsa-miR-92a-3p#sense#chrX;hsa-mir-92a-2,48,70

and it is extracting the precursor as: chrX;hsa-mir-92a-2

I guess sRNAbench is was run with some other flag, can you pass me you command line used on that step?

Thanks!

kicheolkim commented 4 years ago

I used this command: java -jar $sRNApath/sRNAbench.jar input=$inputDir/$fileName output=$outputDir/"$sample" p=$thread dbPath=$dbPath adapter=$adapter minReadLength=15 holdNonAdapter=true umi=3pA12 qualityType=mean minQ=20 microRNA=hsa mature=mature.fa hairpin=hairpin.fa species=hg38 alignType=n seed=18 noMM=1 predict=false isoMiR=$isomir

sRNAbench has two modes: library mapping and genome mapping. I used genome mapping mode in this output, I think that's why it has a chromosome tag. It works well with output from library mode.

Thanks!!

miRTop / mirtop