RNAseq mapping with Salmon for differential expression

How do you process raw read data for the purpose of differential expression? In this video I map raw RNAseq reads using Salmon and follow up with differential expression analysis in R with Deseq2.
notebook:
github.com/mou...
references:
salmon.readthe...
index preparation:
combine-lab.gi...

[@ziqifu2232]

This is very helpful. Thank you!

[@lst595991]

I really like the names of these tools. For example, the package called "Wasabi" can prepare output from Sailfish and Salmon for downstream analysis with the program "sleuth," and its function is named "prepare_fish_for_sleuth."
As for "Kallisto," I'm not sure about the reason behind its name.

[@sanbomics]

haha yeah people generally do a good job with naming. Or maybe only the well named tools become popular... hmmm

[@aereoalex]

Thanks for generously sharing your knowledge. I wonder if you could make a video showing how to use MUSIC2 for deconvolution of bulkRNAseq in 2 conditions using public SingleCell RNAseq,
CHEERS,

[@gregj3913]

Thank you for this video!
I am confused why you can use the Ensembl version 109 for TxImport--but you ran salmon with the Gencode transcripts fasta. Doesn't gencode differ in what transcripts are included versus Ensembl? Or this doesn't really matter?

[@Daoro123]

Thank you for this video. I have a question: I'm don't really understand the need to get a decoy. I mean why not map directly to the full transcriptome??

[@sanbomics]

A decoy will increase specificity because reads that map better to the decoy than the transcripts shouldn't be included in your counts. Reads can have multiple alignments

[@blackmatti86]

Hi, just a question, do you run any bioinformatics courses or do you offer 1:1 tuition/mentorships? Despite the great wealth of information and tutorials you share, I sometimes struggle to follow them using my own samples. Thank you!

[@sanbomics]

I don't, but I would offer it with my normal consulting rate of 125/hr.

[@benaawf9227]

Thank you ......I've assembled a new transcriptome from a non-model organism using Trinity. To functionally annotate it, I used Diamond_blastx against the SwissProt database. However, I'm stuck on generating tx2gene since there's no reference genome for my species in Ensembl or Genecode. Could you direct me to a resource or tutorial that can help me generate gene IDs and transcripts for my newly assembled non-model organism?

[@Brickkzz]

Thanks for another gem Mark (shame it's not Python though 😂).
This might be a long shot but are you aware of any established tools for determining dominant transcriptional isoforms in the sequenced samples?

[@sanbomics]

Well you can export the counts and open them in python xD. Salmon does that! In this video I sum all the transcripts for a given gene. But the raw quant.sf file gives you the estimated counts per transcript/isoform.

[@user-jd1zh3rk5w]

hi, when i try to use salmon index command, it says "[unrecognised option '-d']. how can i solve this ? Thank you.

[@vjsanchezarevalo]

Thanks! Is there any alternative to tximport in Python to use with pydeseq2?

[@sanbomics]

Hi! Not that I am aware of. For now, just export the counts as a csv immediately after creating dds

[@jeffk8900]

Thanks!

[@sanbomics]

Ohhh, thank you! Really appreciate the tip! :)

[@scheket]

Hi, can’t build the indexes for mouse( always receive “killed” after some time(

[@sanbomics]

Hmm, are you running out of memory? Try building it without decoys as a last restort