Modules

StrainCascade v2.0.0 comprises 30 analysis modules (SC1–SC30), each operating within its own Apptainer container for full reproducibility and isolation.

Module Reference

Module Tool Description
SC1 Canu Long-read correction & trimming
SC2 LJA Multiplex de Bruijn graph assembly (HiFi-optimised)
SC3 SPAdes Multi-k-mer de Bruijn graph assembly
SC4 Canu Overlap-layout-consensus assembly with adaptive k-mer weighting
SC5 Flye Repeat graph assembly for long error-prone reads
SC6 Unicycler Hybrid (long+short) or long-read-only assembly (miniasm+Racon)
SC7 QUAST Assembly quality evaluation round 1 (pre-merge)
SC8 MAC2 Consensus assembly merging from multiple assemblers
SC9 QUAST Assembly quality evaluation round 2 (post-merge)
SC10 Circlator Contig circularisation
SC11 QUAST Assembly quality evaluation round 3 (post-circularisation)
SC12 Arrow/Medaka + Polypolish Assembly error correction (long-read + optional short-read polishing)
SC13 minimap2 + BBMap Read mapping & coverage statistics
SC14 CheckM2 Genome completeness & contamination QC
SC15 GTDB-Tk Taxonomic classification
SC16 GTDB-Tk De novo phylogenetic tree construction
SC17 Bakta Genome annotation (v2)
SC18 Prokka Prokaryotic genome annotation
SC19 DeepFRI Deep-learning protein function prediction (GO + EC)
SC20 MicrobeAnnotator Metabolic pathway annotation
SC21 PlasmidFinder Plasmid replicon identification
SC22 AMRFinderPlus Antimicrobial resistance gene detection
SC23 ResFinder Resistance gene identification
SC24 dbCAN3 Carbohydrate-active enzyme (CAZyme) identification
SC25 IslandPath-DIMOB Genomic island prediction
SC26 VirSorter2 Viral/phage sequence detection (HMM multi-classifier)
SC27 geNomad Viral & plasmid identification (marker-based classification)
SC28 CRISPRCasFinder CRISPR-Cas system detection
SC29 ISEScan Insertion sequence element detection
SC30 — Consolidated data integration & interactive HTML report

Module Groups

Assembly (SC1–SC8)

Five distinct assembler algorithms maximise complementarity:

  • Canu — adaptive k-mer weighting with overlap-layout-consensus
  • SPAdes — multi-k-mer de Bruijn graph approach
  • Flye — repeat graph construction for long error-prone reads
  • LJA — multiplex de Bruijn graphs optimised for HiFi reads
  • Unicycler — hybrid mode (long+short) or long-read-only (miniasm+Racon)

Draft assemblies are evaluated by QUAST (SC7), then merged into a consensus assembly by MAC2 (SC8).

Refinement (SC9–SC13)

Post-merge, the best assembly undergoes:

  1. Quality evaluation (SC9) — QUAST assessment post-merge
  2. Circularisation (SC10) — Circlator resolves circular replicons
  3. Quality evaluation (SC11) — QUAST assessment post-circularisation
  4. Polishing (SC12) — Arrow (PacBio) or Medaka (ONT) for long-read polishing, plus optional Polypolish when short reads are provided
  5. Coverage (SC13) — minimap2 alignment + BBMap statistics

Quality Control & Taxonomy (SC14–SC16)

  • CheckM2 (SC14) — completeness/contamination assessment using ML-based marker gene analysis
  • GTDB-Tk (SC15) — taxonomic classification against the Genome Taxonomy Database
  • GTDB-Tk (SC16) — de novo phylogenetic tree placement

Annotation (SC17–SC20)

A multi-tool annotation consensus framework:

  • Bakta (SC17) — primary annotation (NCBI-compliant, database-driven)
  • Prokka (SC18) — complementary gene calls
  • DeepFRI (SC19) — deep-learning protein function prediction across GO categories (MF, BP, CC) and EC numbers
  • MicrobeAnnotator (SC20) — metabolic pathway annotation via KEGG/COG

When Bakta and Prokka disagree, DeepFRI serves as an equal voting member in a three-way consensus. If DeepFRI agrees with one tool, that annotation is preferred; when all three disagree, structure-based predictions are considered alongside database-derived annotations.

Functional Screens (SC21–SC29)

Module Target Method
SC21 Plasmid replicons PlasmidFinder database matching
SC22 AMR genes AMRFinderPlus HMM + BLAST
SC23 Resistance genes ResFinder database screening
SC24 CAZymes dbCAN3 multi-tool consensus
SC25 Genomic islands IslandPath-DIMOB dinucleotide bias
SC26 Phages/viruses VirSorter2 HMM multi-classifier
SC27 Viruses & plasmids geNomad marker-based classification
SC28 CRISPR-Cas CRISPRCasFinder pattern matching
SC29 IS elements ISEScan profile HMM scanning

Data Integration (SC30)

All module outputs are consolidated into a unified results directory with an interactive HTML report for visual exploration.

Execution Modes

Mode Modules Use Case
minimal SC3, SC7, SC14, SC15, SC17, SC30 Quick assembly check
efficient SC1–SC3, SC7–SC12, SC14–SC15, SC17, SC21, SC30 Balanced analysis
standard SC1–SC6, SC7–SC15, SC17–SC30 Comprehensive analysis (default)
comprehensive All 30 modules Full analysis
custom User-specified (e.g. custom:SC1,SC2,SC3) Targeted analysis

Bundles

Bundle Focus Modules
assembly Genome assembly & QC SC1–SC14
annotation Annotation modules SC17–SC20
functional Functional screens SC15, SC22–SC25
phage Phage detection SC26–SC29

Assembly-Input Mode

When pre-assembled genomes are provided via -a, the following modules are automatically skipped:

  • SC1 (Canu trim), SC2–SC6 (assemblers), SC7–SC8 (evaluation 1 + merge), SC9 (evaluation 2), SC10 (circularisation), SC12 (polishing), SC13 (coverage)

The pipeline starts from SC11 (evaluation 3) and SC14 (CheckM2) onwards.

Back to top