Providing multiple read files to Salmon

Question

[ExtentHonest56]

In Salmon's documentation, they state you can provide multiple read files when "one may wish to quantify multiple replicates or samples together, treating them as if they are one library." If, for example, you input 2 separate fasta files from the same sample (collected on separate days) will the output will be as if you submitted only 1 sample? Or am I interpreting this incorrectly?

[GenoMax]

same sample (collected on separate days)

How can the same sample be collected on separate days? Did you prepare a biological replicate on separate days?

[ExtentHonest56]

Yes, apologies. Same sample source (same animal) but collected twice on two separate days, exposed to different conditions on each day.

[GenoMax]

Thanks for clarifying.

[Lada]

hi,

I am investigating the same question at the moment: my input is -1 replica1_read1 replica2_read1 replica3_read1 -2 replica1_read2 replica2_read2 replica3_read2 and I got :

• in the quant.sf file just one column for TPM
  • in aux_info/meta_info.json i got "num_libraries": 1 and "num_processed" reads matches total number of reads (sum for all three samples) I hope there is a way to treat this as 3 different libraries in one script so i don't have to run 3 separate analysis.

[Trivas]

From the documentation:

Providing multiple read files to Salmon

Often, a single library may be split into multiple FASTA/Q files. Also, sometimes one may wish to quantify multiple replicates or samples together, treating them as if they are one library.

So your output is correct. Unfortunately, I think your best bet is to run 3 individual commands which IMO is good practice especially if you want to switch into parallelization.

[Lada]

yeah, I figured the same. tnx for the confirmation :)

Answer 1

[Trivas]

Think of it as they merge the FASTQ files together, then run Salmon on the merged files. So it will be as if you submitted only oen sample.

[ExtentHonest56]

Hmm that is what I was interpreting. This is interesting to me. I have samples collected on the same sources twice, from 2 different "groups" on two different days (exposed to separate conditions on each day), and was thinking this might be an interesting way to view differences between the groups (disregarding day/conditions) to get a broader view. Just exploring my data and options, thank you!