Skip to content

Installation

HelixForge requires Python 3.10 or newer.

Terminal window
pip install helixforge

The console script helixforge is installed on your PATH. The command-line interface (click) and result rendering (rich) are part of the base install, so a bare pip install helixforge gives you a working CLI.

Optional feature sets, installed with pip install "helixforge[<extra>]":

ExtraAdds
clicommand-line interface (click, rich)
statstidy stats tables (pandas)
vizplotting and locus figures (matplotlib, plotly, dna-features-viewer)
bigwig, tracksbigWig coverage support (pyBigWig)
hpcHyperShell distributed execution backend
devtest and type-check baseline

HelixForge drives several bioinformatics tools as subprocesses. They are not Python dependencies and must be installed separately (conda or container):

  • Mikado, the reconciliation engine
  • TransDecoder, ORF calling for the backstop
  • DIAMOND, protein homology search
  • Portcullis, splice-junction filtering
  • miniprot, protein-to-genome alignment
  • StringTie, reference-free transcript assembly
  • STAR, RNA-seq read alignment
  • samtools, BAM/CRAM handling

Verify the whole environment before a long run:

Terminal window
helixforge doctor
  1. It resolves and versions every external tool the pipeline shells out to (mikado, diamond, transdecoder, portcullis, …) and reports the path and detected version.

  2. It checks environment hygiene, the console-script shim and CRAM reference setup.

  3. Given --genome + --helixer (and optional evidence), it runs the input-integrity preflight: seqid concordance across inputs, GFF3/GTF/FASTA/BAM format sniffing, and BAM/CRAM index presence.

It exits non-zero if a required tool is missing or a gate fails, so a misconfigured environment fails in seconds rather than hours. Nothing is modified, doctor only inspects. See the doctor CLI reference for all options, and the preflight tutorial step for the maize invocation.