name: salmon_partial_sa_index
version: 0.0.1
output_asset_class: salmon_partial_sa_index
description: |
  Transcriptome index for salmon, produced with salmon index using partial
  selective alignment method. Preparation includes transcriptome mapping to the genome
  and extraction of the relevant portion out from the genome and indexing it along
  with the transcriptome. Recipe source -- https://github.com/COMBINE-lab/SalmonTools/blob/master/scripts/generateDecoyTranscriptome.sh
input_files: {}
input_params:
  threads:
    default: "8"
    description: Number of threads to use for parallel computing
  kmer:
    default: "31"
    description: The length of kmer to use to create the indices
input_assets:
  fasta:
    asset_class: fasta
    default: fasta
    description: fasta asset for genome
  fasta_txome:
    asset_class: fasta
    default: fasta_txome
    description: fasta asset for transcriptome
  ensembl_gtf:
    asset_class: gtf
    default: ensembl_gtf
    description: GTF file for exonic features extraction
docker_image: combinelab/salmon
command_templates:
  - 'gunzip -c {{values.genome_folder}}/{{values.assets["ensembl_gtf"].seek_keys_dict["gtf"]}} > {{values.output_folder}}/{{values.genome_digest}}.gtf'
  - "awk -v OFS='\t' '{if ($3==\"exon\") {print $1,$4,$5}}' {{values.output_folder}}/{{values.genome_digest}}.gtf > {{values.output_folder}}/exons.bed"
  - 'bedtools maskfasta -fi {{values.genome_folder}}/{{values.assets["fasta"].seek_keys_dict["fasta"]}} -bed {{values.output_folder}}/exons.bed -fo {{values.output_folder}}/reference.masked.genome.fa'
  - 'mashmap -r {{values.output_folder}}/reference.masked.genome.fa -q {{values.genome_folder}}/{{values.assets["fasta_txome"].seek_keys_dict["fasta"]}} -t {{values.params["threads"]}} --pi 80 -s 500 -o {{values.output_folder}}/mashmap.out'
  - "awk -v OFS='\t' '{print $6,$8,$9}' {{values.output_folder}}/mashmap.out | sort -k1,1 -k2,2n - > {{values.output_folder}}/genome_found.sorted.bed"
  - "bedtools merge -i {{values.output_folder}}/genome_found.sorted.bed > {{values.output_folder}}/genome_found_merged.bed"
  - "bedtools getfasta -fi {{values.output_folder}}/reference.masked.genome.fa -bed {{values.output_folder}}/genome_found_merged.bed -fo {{values.output_folder}}/genome_found.fa"
  - "awk '{a=$0; getline; split(a, b, \":\"); r[b[1]] = r[b[1]]\"\"$0} END { for (k in r) { print k\"\\n\"r[k] }}' {{values.output_folder}}/genome_found.fa > {{values.output_folder}}/decoy.fa"
  - 'cat {{values.genome_folder}}/{{values.assets["fasta_txome"].seek_keys_dict["fasta"]}} {{values.output_folder}}/decoy.fa > {{values.output_folder}}/gentrome.fa'
  - "grep '>' {{values.output_folder}}/decoy.fa | awk '{print substr($1,2); }' > {{values.output_folder}}/decoys.txt"
  - "rm {{values.output_folder}}/exons.bed {{values.output_folder}}/reference.masked.genome.fa {{values.output_folder}}/mashmap.out {{values.output_folder}}/genome_found.sorted.bed {{values.output_folder}}/genome_found_merged.bed {{values.output_folder}}/genome_found.fa {{values.output_folder}}/decoy.fa {{values.output_folder}}/reference.masked.genome.fa.fai"
  - 'salmon index -t {{values.output_folder}}/gentrome.fa -d {{values.output_folder}}/decoys.txt -i {{values.output_folder}} -k {{values.params["kmer"]}} -p {{values.params["threads"]}}'
custom_seek_keys:
  version: "salmon --version | awk '{print $2}'"
default_asset: "{{values.custom_seek_keys.version}}"
tags:
- quantification
- rna-seq
- pseudoalignment
- selective-alignment
outputs:
- pattern: '*.json'
  description: Salmon index metadata files
- pattern: seq.bin
  description: Salmon sequence binary
- pattern: pos.bin
  description: Salmon position binary
- pattern: gentrome.fa
  description: Combined genome and transcriptome FASTA
resources:
  memory: 16GB
  disk: 30GB
  time: 4h
test:
  commands:
  - test -d {output_dir}
  - test -f {output_dir}/seq.bin
metadata:
  author: nsheff
  created: '2026-05-20'
  license: BSD-2-Clause