Quick start

On the toy example

The repository includes a toy transcriptome with an exon skip, an alternative 3′ splice site and a pair of paralogs. Simulate reads from it (with simulate, a developer command not listed in --help), then index and map:

spearfish simulate fixtures/toy/transcripts.fa -n 1000 -o reads.fa --truth truth.tsv
spearfish index fixtures/toy/transcripts.fa -o toy.sfi
spearfish map toy.sfi reads.fa -o map.tsv

The simulated read names carry their origin, and truth.tsv the true counts.

map.tsv has one line per reported transcript for each read:

read                    read_len strand rank transcript  score  qstart qend tstart tend anchors anchored max_disc flags
r0|TOYT0000003|-|91|1100     997 -      0    TOYT0000003 344.00 0      997  91     1100 11      371      4        -
r0|TOYT0000003|-|91|1100     997 -      1    TOYT0000004 253.50 11     868  221    1053 10      321      31       anchor_span

Next to it, map.tsv.meta.json records the run: the spearfish build, the command line, the index, every parameter used and the read counts (Output).

On real data

Index a transcriptome once, then map each sample with the preset for its read type:

spearfish index gencode.v50.transcripts.fa.gz -o gencode_v50.sfi -t 16
spearfish map gencode_v50.sfi sample.fastq.gz --preset ont -t 16 -o sample.tsv

Use --preset drna for ONT direct RNA and --preset hifi for PacBio HiFi (Mapping reads). Add --paf sample.paf for PAF output, or --rad sample.rad for oarfish.

For transcript abundances, quantify the RAD file with oarfish, or let oarfish map with spearfish itself:

spearfish map gencode_v50.sfi sample.fastq.gz --preset ont -t 16 --rad sample.rad
oarfish -a sample.rad -j 16 -o sample

This writes oarfish’s estimates to sample.quant (Quantification).