Could I use BAM as input rather than FASTQ?

[juheon]

I know the current version of SQANTI2 does not support BAM as input. It uses FASTQ files and runs the alignment. If I want to do some post-alignment filtering before SQANTI2, what I need to do is 1) align FLNC reads on my own, 2) do post-alignment filtering, 3) convert bam to FASTQ, 4) run SQANTI2 which comes with an additional round of alignment using filtered FASTQ. This is more issue since I use FLNC reads rather clustered reads and I need more post-processing for FLNC reads. For my research, the number of PacBio reads matters to assess the expression of isoforms. Moreover, the clustering procedure discards singleton transcripts which is the evidence of expression of rare isoforms. Would it be possible to use BAM as input for SQANTI2?

[Magdoll]

Hi @juheon ,

I will put BAM/SAM support as a possible future support.

In the meantime, you can convert the SAM/BAM alignment into GTF/GFF3 and use --gtf as input. This way, SQANTI2 will not run aligner again.

--Liz