Star alingment, Salmon quantification

Question

[lidiaryabova]

Hello, I hope you all well in these times. I am very new in bioinformatics, currently doing pipeline with human diseased and control samples using STAR aligner and I got as an output .bam files. I need to put .bam files into Salmon for quantification and can someone help me to understand how it is done, please? If you can suggest any existing pipeline that I can have a look at how it is done? I generated Salmon index code : salmon index \ -t /Homo_sapiens.GRCh38.cdna.all.fa \ -i SalmonIndex \ -k 31 will appreciate for any advice in STAR-SALMON !

[GenoMax]

lidiaryabova : See the link https://salmon.readthedocs.io/en/latest/salmon.html#quantifying-in-alignment-based-mode

Answer 1

[Rob]

You can use salmon with STAR aligned reads, but you must tell STAR to prepare a bam with respect to the transcriptome (I believe the flag is --quantMode TranscriptomeBam, and probably some other flags you want too. In fact STAR+salmon is a nf-core workflow (https://github.com/nf-core/rnaseq) so you can check there for all the details, and maybe use nf-core if it suits your needs. Finally, as Michael points out, you could also just have salmon align the reads itself via selective alignment.

[lidiaryabova]

Hello, thank you for your advice, I did run STAR quite well in 'diseased ' and 'control ' samples , but I have mapping rate 97% in control and 94% in the "diseased sample" .I am not sure it is good ? Could you share your expirience with mapping using STAR tool or maping expirience with any other tool . Is 97% is good? I mean compare to 94% it is sounds good .I think th error was- 'samples are too short'. My samples are 100 in lenght and I use flag . But how I can make it better please? I was thinking about double mapping using STAR , but never I used it before ... If you know how does it work please share your expireence and your opinion please ...Thank you in advance .