Transcriptome alignment about nanoseq HOT 6 CLOSED

Having a brainstorm and have come up with a few possible ways to do this:

1. Have a --align_transcriptome flag or similar and when this is activated assume all reference fasta files are transcriptomes

2. Allow transcriptome to be specified in the sample sheet by entering transcriptome path/to/ref.fa in the genome section i.e.

sample,fastq,barcode,genome
K562_RUN1_REP3,,1,transcriptome path/to/ref.fa
HEPG2_RUN3_REP5,,2,path/genome.fa
CALIBRATION_RUN,,3,

In this case we could parse the transcriptome section to understand which references are transcriptomes and which are genomes.

3. Have a 5th column in the samplesheet for gtf/transcriptome(fa)

sample,fastq,barcode,genome,transcript
K562_RUN1_REP3,,1,,path/to/ref.fa
HEPG2_RUN3_REP5,,2,path/genome.fa, path/to/annot.gtf
CALIBRATION_RUN,,3,

In this case the 5th column would be optional - logic would be as follows

if 5th col exists:
     if gtf:
          check the genome file exists
          if minimap2:
               convert to bed12
               perform transcript aware genome alignment using `--junc-bed` flag
          if graphmap2:
               perform transcript aware genome alignment using `--gtf` flag
     else if fa:
          perform transcriptome alignment

Keen to get an agreement on how we should implement this as well as #31 so we can start developing.

from nanoseq.

This is a tricky one especially since we could have the possibility where we have different genomes/transcriptomes for the samples in the samplesheet.

I think we should just have an additional transcriptome entry in the samplesheet. This could either be a fasta transcriptome or a gtf file which we can use to extract the transcripts from the genome fasta file. That's the most flexible option.

• If genome is present and not transcriptome map to that. If it's an iGenomes reference then we get the gtf automatically and generate the transcriptome on the fly.
• If genome and transcriptome are present then use transcriptome but will have to make sure transcriptome is a fasta and not gtf.
• If genome isn't present and transcriptome is then use transcriptome.

Working out and validating whether we need to use gtf or fasta for transcriptome will involve quite a bit of refactoring I suspect.

How does that sound?

from nanoseq.

I think that sounds good!
I also feel like it's gonna require a lot of refactoring but it will give a lot of flexibility and functionality to our pipeline that will be worth it.
Especially for downstream steps like nanopolish etc. where transcriptome alignment is required.

from nanoseq.

Fixed in #46

from nanoseq.

Some things left to do:

• There still may be some bugs in the logic so it will need extensive testing with different entries for genome and transcriptome, and by using the different --skip flags to see if the channels are all defined properly.
• Add detailed documentation.

from nanoseq.

Additional tests have been added to GitHub Actions to cater for the the testing. Extensive documentation was also added in #57

from nanoseq.