Aaron O'Brien

Finds ITS1, 5.8S and ITS2 inside Nanopore and PacBio amplicons by chaining four profile-HMM anchors, and falls back to two-anchor pairs when a read is missing a flank, so every region it returns still has a model boundary on both sides rather than a read edge. Failures come back as structured codes and a per-sample QC summary instead of a silently shorter FASTA.

Full ITS recovered from a 54,659-read Nanopore library, at 4.6× the speed of ITSx.

Reads that map equally well to several UNITE references are the reason species-level abundance tables lie. EMITS runs expectation-maximization over minimap2 alignments to split those reads probabilistically and aggregate across redundant accessions, with presets for Nanopore and PacBio chemistries.

Pulls small-subunit rRNA out of environmental reads on either strand and sorts what it finds by origin, which matters when host or organellar sequence would otherwise swamp the target community.

Where a sample came from, what was done to it and which run it ended up in — with role-based access so the people doing the bench work can enter their own data.

Metal balances and recovery curves across minicolumn campaigns, deployed on site so results are current rather than reconstructed from spreadsheets at the end.

Snakemake and Nextflow workflows for the centre's Nanopore runs, with PICRUSt2 functional profiling and interactive D3 reports for vineyard soil, bioleaching and acid mine drainage, compost time series and food-safety surveillance.