Minimap2 configuration for direct RNA-seq

[mfurla]

Dear IsoQuant developers,

Thank you very much for this great tool!

I would like to include IsoQuant in my pipeline for direct RNA-seq data analysis and, to avoid wasting computational resources, I would like to run the tool on previously aligned BAM files.

In this regard, could you please tell me the recommended minimap2 parameters for IsoQuant?

In your manuscript, you report the typical configuration: -ax splice -uf -k14, which, however, is not identical to the one used by the latest version of IsoQuant (according to the header of a BAM produced by the tool): -ax splice --secondary=yes -Y --MD -uf --junc-bed /path/to/bed /path/to/k14_idx.

Any insight on this would be greatly appreciated.

Best,

Mattia

[andrewprzh]

Dear @mfurla

The minimap2 configuration has not changed since the publication, -Y simple reports sequences for secondary alignments and --junc-bed /path/to/bed /path/to/k14_idx provides the annotation. I think these options are more or less optimal. At least this is what IsoQuant expects.

IsoQuant can be ran with previously used BAM via --bam. Also, if the BAM file was generated during IsoQuant run, next time IsoQuant should find it automatically if the same FASTQ files is provided.

Best
Andrey

[mfurla]

Dear Andrey,

Thank you very much for your feedback.

I have a follow-up question regarding secondary alignments. It is not clear to me whether they are required for a standard IsoQuant run or can be safely discarded. Indeed, they are retained in the BAM file according to the alignment parameters, but in a previous issue you mentioned that, starting from version 3.9, they are no longer used by default.

This could be relevant for speeding up IsoQuant execution and for maintaining consistency with tools such as modkit. This point was also mentioned in another issue, but it was not further discussed.

Thank you again for your help!

Best regards,

Mattia

[andrewprzh]

Dear @mfurla

Yes, you are absolutely right. Secondary alignments are not used by default anymore, but minimap still reports them (in case a user would like to use them or change IsoQuant option without rerunning minimap).
At some point, I will do some experiments and see if changing minimap options will give a noticeable speed up.

Best
Andrey

[mfurla]

Dear Andrey,

Thank you very much for your feedback!

Based on your last message, I decided to run IsoQuant on BAM files containing primary and supplementary high-quality alignments (-q 20).

In this regard, I was wondering whether supplementary alignments are processed by IsoQuant.

Best,

Mattia

[andrewprzh]

@mfurla

No, current version ignores supplementary alignment.

Best
Andrey