Salmon error message

Question

[saadleeshehreen]

I am trying to quantify the expression of my genes by using salmon. The command is following

salmon quant -p 10 -i ensembl_index/ -l A -1 SRR1056045_2.fastq.gz -2 SRR1056045_1.fastq.gz -o ensembl_quant

But got the following error message:

Error reading from the FASTA/Q stream. Minimum return code for left and right read was (-2). Make sure the file is valid.

I was trying to analyse 5 .sra files which were downloaded from ncbi. Two of my files were OK, but enter-count problems with rest three. Same error message for them. Is it the problem with those .sra files?

Thanks

[jomo018]

If you are worried about sra files, you can directly download your fastq.gz files here:

https://www.ebi.ac.uk/ena/data/view/SRR1056045

[mr.udvr]

Dear Biostars Community

salmon throws me the exact same error message "Error reading from the FASTA/Q stream. Minimum return code for left and right read was (-2)". However I have my own files, and I get this error only for one file out of 20.

This is the code I use:

#!/bin/bash
io="/storageNGS/ngs3/projects/other/Example"
output="salmon_outputs"
mkdir -p $output

nr_threads=42
[ -z "$2" ] || nr_threads=$2

for i in `ls -1 $io/mergedFASTQ/$1*.fastq | sort | awk '{if (NR%2==1) print $0}'`
do
  j=`echo $i | sed s/read1concat/read2concat/`
  sample_name=`basename $i "-read1concat.fastq"`
  salmon quant -i mm9_rna_salmon_index -l A -1 $i -2 $j -p $nr_threads --validateMappings -o $output/$sample_name
done

All my files are named *-read1concat.fastq and *-read2concat.fastq. For 19 it's fine, for one is not. What could be wrong with my file? I already used this file before, and had no problems. Thank you for any hint!

[ATpoint]

Plese do not ask questions in old threads. Open a new one. Go manually through your code for one representative iteration and see if $i and $j give the necessary output. Probably something is flawed parsing the two files in your script.

[Rob]

Salmon's parser uses kseq under the hood, and it simply propagates the kseq error code if it encounters one. This one seems to suggest that one of the files had a truncated quality string (i.e. the quality string was shorter than the corresponding read).

Answer 1

[Devon Ryan]

A couple things:

1. One of those files (likely SRR1056045_1.fastq.gz) is corrupt. Just download it again.
2. You flipped which file goes where. The _1 file should be with -1 and the _2 file should be with -2. Granted, salmon may figure out you've swapped things since you specified -l A, but there's no need to chance that.