Quick start

This walks through the two-command workflow: build an index from a transcriptome, then quantify a paired-end sample.

1. Build an index from your transcriptome FASTA.

   salmon index \
     -t transcripts.fa \
     -i salmon_index \
     -p 16

   The index is reusable across every sample quantified against the same transcriptome. For decoy-aware indexing, append the genome to the FASTA and pass -d decoys.txt (one decoy sequence name per line; decoy records must come last in the FASTA).

2. Quantify a paired-end sample. -l A auto-detects the library type.

   salmon quant \
     -i salmon_index \
     -l A \
     -1 reads_1.fastq.gz \
     -2 reads_2.fastq.gz \
     -p 16 \
     -o sample1_quant

   For single-end reads, use -r reads.fastq.gz instead of -1/-2.

3. Read the results. sample1_quant/quant.sf has one row per transcript:

   Name        Length  EffectiveLength  TPM         NumReads
   NM_004503   1681    1502.723         36410.66    331.969
   ...

   Load it in R with tximport:

   library(tximport)
   txi <- tximport("sample1_quant/quant.sf", type = "salmon", txOut = TRUE)

Common options

• Faster, alignment-free mapping: add --sketch. See mapping modes.
• Bias correction: --seqBias, --gcBias, --posBias.
• Posterior uncertainty: --numBootstraps 100 or --numGibbsSamples 100. See inferential replicates.
• Gene-level output: -g txp2gene.gtf also writes quant.genes.sf.

See the CLI reference for the full option list.