Linclust2, Cluster2 & ProteomeSearch

[Gyuuul2]

Overview

Default workflow setting: Linclust2 (with alignmentMode 3), Cluster1
Regression tests need to be updated accordingly.

Module Updates

• kmermatcher: multithreading support, fixed inconsistency in split mode, added adjacency-based and count-table-based center swapping (spaced kmer is not included)
• align2clust: new module for alignment-based clustering
• BlockAligner: added banded block aligner for ungapped/gapped alignment in align2clust
• proteomecluster: updated workflow
• proteomesearch / parseproteomealignment: new modules

Workflow Updates

• Added --linclust-version {1,2} and --cluster-version {1,2} parameters
• Added clusthash as an optional first step in the linclust workflow
• Linclust2: kmermatcher → align2clust
• Cluster2: linclust2 → prefilter → align2clust
  • During merging, a cluster is only merged if ungapped alignments between the representative and all members of the newly merged cluster pass the filtering criteria