Classification CLI Workflow

CLI Workflow

Map reads to a reference genome

Map reads using pbmm2.

pbmm2 align --preset ISOSEQ --sort <input.bam> <ref.fa> <mapped.bam>

Collapse into unique isoforms

Collapse redundant transcripts into unique isoforms based on exonic structures using isoseq collapse.

Single-cell IsoSeq:

isoseq collapse <mapped.bam> <collapsed.gff>

Bulk IsoSeq:

isoseq collapse --do-not-collapse-extra-5exons <mapped.bam> <collapsed.gff>

Sort input transcript GFF

Sort the transcript GFF file output from isoseq collapse.

pigeon sort <collapsed.gff> -o sorted.gff

Sort and index the reference files

Sort and index the genome annotation, (optional) CAGE peak, and (optional) intropolis files before classification. Sorting prior to indexing ensures that all records for a given chromosome/scaffold are contiguous within the file.

pigeon sort gencode.annotation.gtf -o gencode.annotation.sorted.gtf
pigeon index gencode.annotation.sorted.gtf

pigeon sort cage.bed -o cage.sorted.bed
pigeon index cage.sorted.bed

pigeon sort intropolis.tsv -o intropolis.sorted.tsv
pigeon index intropolis.sorted.tsv

Classify Isoforms

Classify isoforms into categories using the base required input.

pigeon classify <sorted.gff> <annotations.gtf> <reference.fa>

Alternatively, classify isoforms using supplemental reference information. Details in pigeon input.

pigeon classify <sorted.gff> <annotations.gtf> <reference.fa> --fl abundance.txt --cage-peak refTSS.bed --poly-a polyA.list

Filter isoforms

Filter isoforms from the classification output.

pigeon filter <classification.txt>

If you want to generate a filtered GFF, you need to also provide the GFF that was used as input to pigeon classify

pigeon filter <classification.txt> --isoforms <sorted.gff>

The expected output consists of:

*.filtered_lite_classification.txt
*.filtered_lite_junctions.txt
*.filtered_lite_reasons.txt
*.sorted.filtered_lite.gff (only if --isoforms is used)

Report gene saturation

Gene saturation can be determined by subsampling the classification output and determining the number of unique genes at each subsample size.

pigeon report <classification.filtered_lite_classification.txt> <saturation.txt>

Make Seurat compatible input

Output files that are compatible with the downstream Seurat analysis package.

pigeon make-seurat --dedup <dedup.fasta> --group <collapse.group.txt> -d <output_dir> <classification.filtered_lite_classification.txt>

The dedup.fasta file is obtained after running isoseq groupdedup or isoseq dedup. The collapse.group.txt file is obtained after running isoseq collapse.

The output will consist of:

Make-seurat output:
<output_dir>/annotated.info.csv
<output_dir>/info.csv
<output_dir>/genes_seurat/barcodes.tsv
<output_dir>/genes_seurat/genes.tsv
<output_dir>/genes_seurat/matrix.mtx
<output_dir>/isoforms_seurat/barcodes.tsv
<output_dir>/isoforms_seurat/genes.tsv
<output_dir>/isoforms_seurat/matrix.mtx