Bams with the same name cause no differences detected in pairwise

[hbeale]

If the manifest has lists bam files with the same name in different directories (e.g. SRR12801019/Aligned.sortedByCoord.out.bam), the corresponding bed files created by splicedice bam_to_junc_bed will all be named _junction_beds/Aligned.sortedByCoord.out.junc.bed and will overwrite each other.

example of failure

Bam manifest used as input to splicedice bam_to_junc_bed

A19     /mnt/output/star_2.7.11b_2024.12.13/SRR12801019/Aligned.sortedByCoord.out.bam   HEK     control
A20     /mnt/output/star_2.7.11b_2024.12.13/SRR12801020/Aligned.sortedByCoord.out.bam   HEK     SUGP1_knockdown
A23     /mnt/output/star_2.7.11b_2024.12.13/SRR12801023/Aligned.sortedByCoord.out.bam   HEK     control
A24     /mnt/output/star_2.7.11b_2024.12.13/SRR12801024/Aligned.sortedByCoord.out.bam   HEK     SUGP1_knockdown

Bed manifest generated by splicedice bam_to_junc_bed; note that all bed files have the same path

A19     /mnt/output/splicedice_2024.12.18_02.10.36/_junction_beds/Aligned.sortedByCoord.out.junc.bed    HEK     control
A20     /mnt/output/splicedice_2024.12.18_02.10.36/_junction_beds/Aligned.sortedByCoord.out.junc.bed    HEK     SUGP1_knockdown
A23     /mnt/output/splicedice_2024.12.18_02.10.36/_junction_beds/Aligned.sortedByCoord.out.junc.bed    HEK     control
A24     /mnt/output/splicedice_2024.12.18_02.10.36/_junction_beds/Aligned.sortedByCoord.out.junc.bed    HEK     SUGP1_knockdown

looking for differences in the pairwise.tsv output from the end of the pipeline

cat pairwise.tsv | grep -v clusterID | cut -f2-7 | sort | uniq | wc -l

1

example of success

create links to bam files containing a unique name and update the manifest

Bam manifest used as input to splicedice bam_to_junc_bed

A19     /mnt/output/star_2.7.11b_2024.12.13/SRR12801019/SRR12801019_Aligned.sortedByCoord.out.bam       HEK     control
A20     /mnt/output/star_2.7.11b_2024.12.13/SRR12801020/SRR12801020_Aligned.sortedByCoord.out.bam       HEK     SUGP1_knockdown
A23     /mnt/output/star_2.7.11b_2024.12.13/SRR12801023/SRR12801023_Aligned.sortedByCoord.out.bam       HEK     control
A24     /mnt/output/star_2.7.11b_2024.12.13/SRR12801024/SRR12801024_Aligned.sortedByCoord.out.bam       HEK     SUGP1_knockdown

Bed manifest generated by splicedice bam_to_junc_bed; note that all bed files are unique

A19     /mnt/output/splicedice_2024.12.19_23.40.49/_junction_beds/SRR12801019_Aligned.sortedByCoord.out.junc.bed       HEK      control
A20     /mnt/output/splicedice_2024.12.19_23.40.49/_junction_beds/SRR12801020_Aligned.sortedByCoord.out.junc.bed       HEK      SUGP1_knockdown
A23     /mnt/output/splicedice_2024.12.19_23.40.49/_junction_beds/SRR12801023_Aligned.sortedByCoord.out.junc.bed       HEK      control
A24     /mnt/output/splicedice_2024.12.19_23.40.49/_junction_beds/SRR12801024_Aligned.sortedByCoord.out.junc.bed       HEK      SUGP1_knockdown

looking for differences in the pairwise.tsv output from the end of the pipeline

cat pairwise.tsv | grep -v clusterID | cut -f2-7 | sort | uniq | wc -l

41222

potential solutions

• we could add the requirement to have unique bam file names to the documentation
• bed output files could be named using the first column of the manifest (e.g. A19) instead of a permutation of the bam file name