Arcadia-Science/noveltree: v2.0.0 (weeping willow)

v2.0.0 - 2026-03-30 Added Zoogle analysis mode (-profile zoogle): end-to-end physicochemical protein distance analysis with time-calibrated gene family trees, phylogenetic correction, Mahalanobis distances, and permutation tests PROTEIN_PROPERTIES, DATE_GENE_FAMILY_TREES, TIME_CALIBRATE_SPECIES_TREE, BUILD_REFERENCE_CHRONOGRAM, ZOOGLE_ANALYSIS modules Centroid-based analysis mode (--ref_species none) for analysis without a reference species Simplified execution mode (-profile simplified): streamlined pipeline for large datasets (adaptive alignment default, no BUSCO, per-species GeneRax EVAL strategy only) Arcadia production profile (-profile arcadia): combines zoogle + AWS Batch Adaptive three-tier alignment routing (--aligner adaptive): MAFFT_TIER1 (≤200 seqs) → WITCH_TIER2 (≤3000) → FAMSA_TIER3 (larger), with automatic fallback between tiers FAMSA aligner option (--aligner famsa) IQ-TREE with automatic FastTree fallback (--iqtree_fasttree_fallback) PARSE_PHYLOHOGS module: per-OG ortholog/paralog inference and HOG membership directly from GeneRax NHX reconciliation (extract_relationships_from_nhx.py) PHYLO_PROFILES module for summarizing gene duplication, loss, and speciation events per species per gene family Optional proteome preprocessing (--preprocess): TransDecoder for transcriptomes, isoform filtering, minimum protein length filtering Samplesheet support for UniProt proteome IDs (UP*) and NCBI genome accessions (GCA_*/GCF_*) as data sources, with automatic download --test_run_mcl flag for optional MCL inflation parameter testing on species with UniProt accessions Dynamic resource allocation for GeneRax, WITCH, IQ-TREE, and MAFFT based on gene family size (sequence count and max length) RENAME_FASTAS process to standardize FASTA filenames before all downstream processes AWS Batch support (-profile awsbatch) with --awsqueue and --awsregion parameters Singularity support (-profile singularity) with automatic Docker-to-Singularity conversion and image caching --test_run flag: restrict analysis to gene families containing all species (fast smoke test) Eukaryote test dataset (6 Opisthokont species) Makefile for Docker image builds, CITATIONS.md, docs/singularity.md New dependencies: treePL, PATHd8, phangorn Changed Main workflow refactored from monolithic main.nf (~500 lines) to 7 focused subworkflows (INPUT_CHECK, PREPARE_INPUTS, INFER_ORTHOGROUPS, INFER_GENE_TREES, RECONCILE_TREES, RECONCILIATION_SUMMARIES, ZOOGLE) Default aligner: witch → adaptive (three-tier routing enabled by default) Default tree method: fasttree → iqtree Default MSA trimmer: none → clipkit min_ungapped_length default: 20 → 50 min_num_spp_per_og default: 4 → 2 min_prop_spp_for_spptree default: 0.25 → 0.50 max_copy_num_spp_tree default: 5 → 10 mcl_inflation default: '1.5,2.0,2.5,3.0,3.5,4.0,4.5,5.0' → '1.5' FastTree arguments updated to evidence-based maximum accuracy settings (-lg -gamma -bionj -pseudo -spr 4 -mlacc 2 -slownni; Zhou et al. 2018) GeneRax container updated to generax_56f3ed0:1.1.3 GeneRax reconciliation model: UndatedDTL → UndatedDL (gene transfers removed) Rare amino acid handling: selenocysteine (U) and pyrrolysine (O) recoded as X before alignment and reconciliation Species name handling standardized: hyphens used throughout (underscores/spaces auto-converted) UniProt annotation retrieval rewritten to use ID Mapping API (replaces bioservices) MCL inflation selection uses only InterPro scoring (OMA removed — Sorensen-Dice incompatible with 1:1 OMA group IDs) Gene family tree dating uses speciation-only calibrations from GeneRax reconciliation with treePL fixed smoothing (smooth=10) and --age_bracket (default ±20%) calibration brackets (replaces congruification) FILTER_ORTHOGROUPS functionality absorbed into ORTHOFINDER_MCL MAFFT_ADAPTIVE renamed to MAFFT_TIER1 for consistency with WITCH_TIER2 and FAMSA_TIER3 Samplesheet redesigned: required columns reduced from 7 to 3 (species, input_data, input_type); all other columns are optional in any order. Column renames for clarity: file → input_data, mode → input_type, uniprot → has_uniprot_ids, mcl_test → include_in_mcl_test, isoform → filter_isoforms, reference → reference_proteome, shallow_db → busco_shallow, broad_db → busco_broad All samplesheet boolean columns standardized to yes/no (previously uniprot used true/false) Removed ORTHOFINDER_PHYLOHOGS module (replaced by PARSE_PHYLOHOGS using GeneRax NHX directly) FILTER_ORTHOGROUPS module (absorbed into ORTHOFINDER_MCL) PMSF two-pass tree inference (iqtree_pmsf.nf) min_num_grp_per_og, max_copy_num_gene_trees, tree_model_pmsf parameters species_tree_prep module OMA annotation collection and scoring from cogeqc analysis bin/protein_annotation.R, bin/resolve_polytomies.R (replaced by Python versions) taxonomy samplesheet column and taxonomy group counting from orthogroup summaries bioservices dependency (replaced by requests for UniProt ID Mapping API) Congruification approach for gene family tree dating (replaced by reconciliation-filtered speciation-only calibrations) Gene transfer (HGT) inference from GeneRax reconciliation (UndatedDTL → UndatedDL) Unused Docker build contexts (orthoxml_phylohogs, prequal, queryup)