Skipping circRNA detection module if Chimeric.junction.out is empty

[farzamani]

I tried both circtools detect v1.3.1 and DCC v0.5.0 using this command in Snakemake:

I only used one sample, paired-end read

circtools detect {input.samplesheet} \
            -mt1 {input.mate1} \
            -mt2 {input.mate2} \
            -D \
            -G \
            -T {threads} \
            -an {params.annotation} \
            -R {params.repeats} \
            -Pi \
            -M \
            -F \
            -fg \
            -Nr 2 1 \
            -A {params.reference} \
            -O {params.outdir} \
            -t {params.tmp} \
            -k

and

DCC {input.samplesheet} \
            -mt1 {input.mate1} \
            -mt2 {input.mate2} \
            -D \
            -G \
            -T {threads} \
            -an {params.annotation} \
            -R {params.repeats} \
            -Pi \
            -M \
            -F \
            -fg \
            -Nr 2 1 \
            -A {params.reference} \
            -O {params.outdir} \
            -t {params.tmp} \
            -k

However, the Chimeric.junction.out in mate1 is empty. Then I got this error message:

Activating conda environment: .snakemake/conda/d33aad9065977b36b220cf4dba3299a3_
Output folder results/circRNAs/circtools/detect/ss_ID13 already exists, reusing
Temporary folder results/circRNAs/circtools/detect/temp/ss_ID13/ already exists, reusing
circtools 1.3.1 started
40 CPU cores available, using 8
WARNING: File results/circRNAs/circtools/ss_ID13/mate1/Chimeric.out.junction is empty!
One of the input junctions files is empty.
Junction files seem empty, skipping circRNA detection module.
circRNA detection skipped due to empty junction files
Filter mode for detected circRNAs enabled without detection module.
Combine with -f or -D.

Additional context
I have been using this tool several times. This tool worked well if the chimeric.junction file is not empty.

I saw a related issue here: dieterich-lab/DCC#8, but it seems the problem still occurs.

Thanks! Any help would be appreciated :)

Update
I used more than one sample, and make a mate1, mate2, and samplesheet file instead of linking it directly to the Chimeric.out.junction

Here is the command:

circtools detect @{input.samplesheet} \
            -mt1 @{input.mate1} \
            -mt2 @{input.mate2} \
            -D \
            -G \
            -T {threads} \
            -an {params.annotation} \
            -R {params.repeats} \
            -Pi \
            -M \
            -F \
            -fg \
            -Nr 2 1 \
            -A {params.reference} \
            -O {params.outdir} \
            -t {params.tmp} \
            -k

And this is the error message:

Activating conda environment: .snakemake/conda/d33aad9065977b36b220cf4dba3299a3_
Output folder results/circRNAs/circtools/detect already exists, reusing
Temporary folder results/circRNAs/circtools/detect/temp already exists, reusing
circtools 1.3.1 started
Input file names have duplicates, add number suffix in input order to output files for distinction
40 CPU cores available, using 8
WARNING: File results/circRNAs/circtools/dip_ID6/mate1/Chimeric.out.junction is empty!
WARNING: File results/circRNAs/circtools/ss_ID13/mate1/Chimeric.out.junction is empty!
One of the input junctions files is empty.
Junction files seem empty, skipping circRNA detection module.
circRNA detection skipped due to empty junction files
Filter mode for detected circRNAs enabled without detection module.
Combine with -f or -D.

[tjakobi]

Sorry for the late reply, but is the file stored on a remote drive that is mounted locally? Circtools only would print this warning if the file size is 0:

os.stat(filename).st_size == 0:

I have seen such issues on Samba file systems, but not on local storage.