Variation Graph Construction

Build individualized genome and transcriptome variation graphs from Exacto variant calls.

This pipeline produces individualized genome and transcriptome variation graphs that encode an individual’s variants alongside the reference sequence. These graphs are useful for downstream variant-aware analyses that benefit from a graph reference instead of a linear one.

Workflow

%%{init: {'securityLevel': 'loose'}}%%
flowchart LR
    DNAV[/"<span style='white-space:nowrap;color:#414a4c'>DNA variants TSV</span>"/] --> BGV(build‑genome‑var‑graph)
    REFG[("<span style='white-space:nowrap;color:#414a4c'>Reference genome FASTA</span>")] --> BGV
    BGV --> GG[/"<span style='white-space:nowrap;color:#414a4c'>Genome variation graph FASTA</span>"/]

    TXS[/"<span style='white-space:nowrap;color:#414a4c'>Transcript model alignments TSV</span>"/] --> BTV(build‑transcriptome‑var‑graph)
    REFG --> BTV
    BTV --> TG[/"<span style='white-space:nowrap;color:#414a4c'>Transcriptome variation graph FASTA</span>"/]

    click BGV href "../cli/build-genome-var-graph.html" "View build-genome-var-graph docs" _self
    click BTV href "../cli/build-transcriptome-var-graph.html" "View build-transcriptome-var-graph docs" _self

    classDef linked text-decoration:underline;
    class BGV,BTV linked

The DNA variants TSV is the output of call-germline-dna-vars or call-somatic-dna-vars. The transcript model alignments TSV is the *_exacto_assembled_transcript_model_alignments.tsv that call-rna-transcript-vars writes (see Variant-aware Transcriptome Assembly).

Genome variation graph

Add germline and/or somatic DNA variants to the reference genome. --only-variant-sequences yes writes only the variant sequences, without the reference contigs:

exacto build-genome-var-graph \
    --variants-tsv-file tumor_somatic_dna_variants.tsv \
    --fasta-file reference_genome.fasta \
    --output-fasta-file tumor_genome_var_graph.fasta \
    --sequence-prefix tumor

Transcriptome variation graph

Build one sequence per assembled transcript from its structure against the reference genome. Each record is named by its assembled_transcript_name:

exacto build-transcriptome-var-graph \
    --transcript-structures-tsv-file rna_variants/tumor_exacto_assembled_transcript_model_alignments.tsv \
    --fasta-file reference_genome.fasta \
    --output-fasta-file tumor_transcriptome_var_graph.fasta