Salmon error (ex. transcript id appears in the BAM head, but was not in the provided FASTA file)

Question

[tonyya21]

Hello. everyone

I am trying to count aligned reads using Salmon in alignment-based mode.

My command is:

salmon quant -t transcript.fa -l A -a aligned_result.bam -o salmon_quant

I am encountering the following error:

Transcript chr8 appeared in the BAM header, but was not in the provided FASTA file.

The preprocessing steps I performed are as follows:

1. Reference genome indexing (by ncbi).
2. Mapping using STAR.
3. Converting SAM to BAM (by samtools).
4. Sorting (Picard SortSam).
5. Checking for duplicate reads (Picard MarkDuplicates).
6. Splitting reads that contain Ns in the CIGAR string (GATK4 SplitNCigarReads).
7. Recalibration (GATK4 BQSR).
8. Sorting (by samtools).
9. Quantifying transcript-level expression using the Salmon algorithm.

I obtained the transcript fasta file from https://www.gencodegenes.org/human/ and used it for analysis. I also tried creating the transcript fasta file using gffread, but the same error persists.

Could this be a problem with the transcript fasta file or the BAM file?

Thank you in advance for your help.

Answer 1

[ATpoint]

You are using different sources of annotation. NCBI genome (which might use chromosome identifiers such as 1,2,3) and GENCODE for the fasta which uses chr1,2,3... Stick with one source.

By the way, steps 4-8 are not necessary to run salmon, and actually you could skip STAR entirely if the only goal is to run salmon, and use salmon for the mapping itself.

[tonyya21]

Oh, that was my mistake. I accidentally used gencode instead of ncbi in step 1. then, I'll proceed by running salmon for the mapping itself. Thank you