Generic solutions for specific problems

Johannes Köster

 

Harpoon 2022

Variant calling

Varlociraptor

species:
  heterozygosity: 0.001
  ploidy:
    male:
      all: 2
      X: 1
      Y: 1
    female:
      all: 2
      X: 2
      Y: 0

samples:
  normal:
    sex: female
    somatic-effective-mutation-rate: 1e-10
  tumor:
    inheritance:
      clonal:
        from: jane
    contamination:
      by: jane
      fraction: 0.1
    somatic-effective-mutation-rate: 1e-6

events:
  germline: "normal:0.5 | normal:1.0"
  somatic_normal: "normal:]0.0,0.5["
  somatic_tumor_low: "normal:0.0 & tumor:]0.0,0.1["
  somatic_tumor_high: "normal:0.0 & tumor:[0.1,1.0]"
  • technology-agnostic (short, long, dna, rna)
  • small and structural variants
  • fully uncertainty aware
  • integrated bias estimation
  • no technical parameters

Input:

  • aligned reads
  • candidate variants
  • scenario

Output:

  • variant calls
  • evidence details
  • posterior probability of events
  • FDR control instead of tuning score thresholds

Genomic variant presentation

Datavzrd

views:
  coding-variants:
    dataset: coding-variants
    render-table:
      columns:
        symbol:
          link-to-url: https://www.ensembl.org/Homo_sapiens/Transcript/Summary?t={feature}
        hgvsp:
          custom-path: resources/varsome-link.js
        vartype:
          display-mode: hidden
        consequence:
          plot:
            heatmap:
              color-scheme: category20
              scale: ordinal
        read-depth:
          plot: 
            ticks:
              scale: linear

Input:

  • set of tables
  • config with
    • relations
    • column rendering specs
    • custom plot specs

Output:

  • interactive, self-contained HTML
  • search & filter
  • export to PDF and Excel

johannes.koester@uni-due.de

Generic solutions for specific problems

By Johannes Köster

Generic solutions for specific problems

Talk at ETIM 2022

  • 426