search

beliveau-lab / PaintSHOP_pipeline

A scalable machine learning pipeline for probe specificity prediction.
MIT License
11 stars 6 forks source link

Problem in probe alignment when running the pipeline using a different probe length range #14

Open JasonJHughes opened 2 years ago

JasonJHughes commented 2 years ago

Dear Beliveau Lab,

When running the PaintSHOP pipeline using the standard example run but changing the length of probes (to 20mers) causes the sam2pairwise to exit due to finding unmapped reads which is a known issue for sam2pairwise.

Even when I add the recommended fix. https://github.com/mlafave/sam2pairwise/issues/1.

Editing the Snakefile rule on extract duplex information from alignment results rule sam_to_pairwise: .... 'samtools view -F 4 {input} | sam2pairwise > {output}'

It causes errors to occur when parsing the alignment results in parse_pairwise. Due to for example getting a value error where the wrong number of items are passed then what the placement implies.

I was wondering if you have a fix for this.

-Jason

JasonJHughes commented 2 years ago

Also, Is there an option to remove the pairwise duplex incase you get no duplexes from samtools? Since for example at shorter probe length ranges, probes might not align with each other for forming duplexes, and this can cause errors in the code leading to your pipeline failing to run.