search

jts / nanopolish

Signal-level algorithms for MinION data
MIT License
568 stars 159 forks source link

Nanopolish with Spliced Reads #376

Closed ch728 closed 6 years ago

ch728 commented 6 years ago

Hi,

I have reads from direct cDNA sequencing that I would like call variants in. I tried using nanopolish, but it threw an error because it detected spliced alignments so It doesn't seem like it's designed to call variants in RNA-Seq data. Are there any work arounds for this? Thanks.

wdecoster commented 6 years ago

I guess an alternative would be to align your reads to the transcriptome, rather than to the genome, and as such avoid spliced alignments. But this gets dodgy when multiple transcripts exist. I don't know what Jared thinks about this, though.

jts commented 6 years ago

Hi,

Sorry but variant calling from spliced reads is not yet supported. I plan to add this feature in the future though so I'll keep this issue open.

Jared

ch728 commented 6 years ago

As a stopgap, I aligned to a transcriptome using minimap2 without splicing but in the post-run summary, it says that all my reads have no alignment and my vcf file is empty. See two sample lines below.

[post-run summary] total reads: 86, unparseable: 0, qc fail: 0, could not calibrate: 0, no alignment: 86, bad fast5: 0 [post-run summary] total reads: 122, unparseable: 0, qc fail: 0, could not calibrate: 0, no alignment: 122, bad fast5: 0

I'm not sure what's going on.

jts commented 6 years ago

This is the same issue we are discussing in #378 so I'm closing this one.