refactor(wdl): strip protein-annotation block from HvpViralPipeline

HvpViralProteinAnnotation is now a standalone Dockstore workflow, so the
nested call inside HvpViralPipeline was redundant — and forced three
optional + two required inputs (assembly_primary_fa, rescue_viral_fa_gz,
bfvd_db_tgz, prostt5_weights_tgz, bfvd_metadata_tsv) that the viral-
prediction half does not consume.

Reverts HvpViralPipeline's input surface to merged_fa_gz + sample_name +
the three viral DBs only. Removed inputs were never produced inside
HvpViralPipeline; they came from HvpAssembly / HvpReadRescue and were
plumbed through purely to feed the nested sub-workflow.

Users wanting protein annotation now invoke HvpViralProteinAnnotation
as a separate Dockstore workflow, sourcing its inputs directly from the
Terra data-table columns populated by HvpAssembly, HvpReadRescue, and
HvpViralPipeline.

miniwdl --strict clean (only pre-existing shellcheck lint in upstream
tasks). pytest 24/24 unchanged.

Author: jonn-smith

wdl/pipelines/PacBio/Metagenomics/HvpViralPipeline.wdl

@@ -4,7 +4,6 @@ import "../../../tasks/Viral/Genomad.wdl"      as GnomadTasks
 import "../../../tasks/Viral/VirSorter2.wdl"   as VS2Tasks
 import "../../../tasks/Viral/CheckV.wdl"        as CheckVTasks
 import "../../../tasks/Viral/ViralSummary.wdl"  as SumTasks
-import "HvpViralProteinAnnotation.wdl"          as ProtAnnot
 
 workflow HvpViralPipeline {
 
@@ -47,39 +46,23 @@ workflow HvpViralPipeline {
             n_provirus:                    "Number of viral contigs classified as provirus",
             n_no_repeats:                  "Number of viral contigs with no terminal repeats",
             mean_completeness:             "Mean CheckV completeness across viral contigs with a completeness estimate",
-            mean_contamination:            "Mean CheckV contamination across viral contigs with a contamination estimate",
-            protein_annotation_combined_faa:    "(Optional, if bfvd_db_tgz + prostt5_weights_tgz supplied) Concatenated AA FASTA from geNomad + Pyrodigal-gv-on-VS2.",
-            protein_annotation_nr_faa:          "(Optional) Non-redundant AA representatives after mmseqs easy-linclust.",
-            protein_annotation_foldseek_hits_tsv: "(Optional) Foldseek alignment hits vs BFVD (BLAST-tab-like TSV).",
-            protein_annotation_best_hits_tsv:   "(Optional) Best foldseek hit per query.",
-            protein_annotation_annotated_tsv:   "(Optional) best_hits left-joined with BFVD metadata when bfvd_metadata_tsv supplied.",
-            protein_annotation_num_nr:          "(Optional) Representative protein count after NR collapse.",
-            protein_annotation_num_hits:        "(Optional) Total foldseek hit-row count.",
-            protein_annotation_num_queries_hit: "(Optional) Distinct queries with at least one foldseek hit.",
-            protein_annotation_num_annotated:   "(Optional) Best-hit queries that joined to a BFVD metadata row."
+            mean_contamination:            "Mean CheckV contamination across viral contigs with a contamination estimate"
         }
     }
 
     parameter_meta {
         merged_fa_gz:       "Merged assembly+rescue FASTA (gzipped) from HvpReadRescue"
-        assembly_primary_fa: "Primary assembly FASTA from HvpAssembly (plain FASTA, uncompressed). Used as a distinct ORF-call source by the optional protein-annotation stage."
-        rescue_viral_fa_gz: "Kraken-flagged viral rescued-reads FASTA (gzipped) from HvpReadRescue. Used as a distinct ORF-call source by the optional protein-annotation stage."
         sample_name:        "Sample identifier used as output file prefix (e.g. bc2097)"
         genomad_db_tgz:     "geNomad database as a single compressed archive (.tar.gz or .tar.zst)"
         vs2_db_tgz:         "VirSorter2 database as a single compressed archive (.tar.gz or .tar.zst)"
         checkv_db_tgz:      "CheckV database as a single compressed archive (.tar.gz or .tar.zst)"
         genomad_extra_args: "Additional command-line args appended verbatim to the genomad invocation"
         vs2_min_length:     "Minimum sequence length for VirSorter2 (default 1000)"
         vs2_extra_args:     "Additional command-line args appended verbatim to the virsorter run invocation"
-        bfvd_db_tgz:         "Optional pre-built BFVD foldseek structure DB (tar.gz, ~534 MB). When supplied together with prostt5_weights_tgz, the HvpViralProteinAnnotation sub-workflow runs end-to-end: Pyrodigal-gv on VS2 viral contigs + geNomad AA -> mmseqs NR -> ProstT5 foldseek DB -> foldseek search vs BFVD -> best-hit + annotation transfer. Leave unset to skip the structural-annotation stage. Source: https://bfvd.steineggerlab.workers.dev/latest/bfvd_foldseekdb.tar.gz"
-        prostt5_weights_tgz: "Optional ProstT5 model weights (tar.gz, ~2.1 GB). Required iff bfvd_db_tgz is supplied. Source: https://foldseek.steineggerlab.workers.dev/prostt5-f16-gguf.tar.gz"
-        bfvd_metadata_tsv:   "Optional BFVD reference metadata TSV (~15 MB). When supplied with bfvd_db_tgz, attaches descriptive annotations to best foldseek hits. Source: https://bfvd.steineggerlab.workers.dev/latest/bfvd_metadata.tsv"
     }
 
     input {
         File   merged_fa_gz
-        File   assembly_primary_fa
-        File   rescue_viral_fa_gz
         String sample_name
 
         File   genomad_db_tgz
@@ -89,12 +72,6 @@ workflow HvpViralPipeline {
         String genomad_extra_args = ""
         Int    vs2_min_length     = 1000
         String vs2_extra_args     = ""
-
-        # Optional structural-annotation stage (HvpViralProteinAnnotation).
-        # Gated on bfvd_db_tgz + prostt5_weights_tgz both being supplied.
-        File?  bfvd_db_tgz
-        File?  prostt5_weights_tgz
-        File?  bfvd_metadata_tsv
     }
 
     # geNomad and VirSorter2 run in parallel — both take the merged assembly+rescue FASTA
@@ -149,21 +126,6 @@ workflow HvpViralPipeline {
             sample_name              = sample_name
     }
 
-    # Optional structural-annotation stage: only runs when both BFVD DB and ProstT5 weights are provided.
-    if (defined(bfvd_db_tgz) && defined(prostt5_weights_tgz)) {
-        call ProtAnnot.HvpViralProteinAnnotation as t_07_ProteinAnnotation {
-            input:
-                sample_name                = sample_name,
-                vs2_viral_combined_fa      = t_02_VirSorter2.viral_combined_fa,
-                genomad_virus_proteins_faa = t_01_Genomad.virus_proteins_faa,
-                assembly_contigs_fa        = assembly_primary_fa,
-                rescued_reads_fa_gz        = rescue_viral_fa_gz,
-                bfvd_db_tgz                = select_first([bfvd_db_tgz]),
-                prostt5_weights_tgz        = select_first([prostt5_weights_tgz]),
-                bfvd_metadata_tsv          = bfvd_metadata_tsv
-        }
-    }
-
     output {
         File  genomad_virus_summary_tsv     = t_01_Genomad.virus_summary_tsv
         File  genomad_virus_fna             = t_01_Genomad.virus_fna
@@ -204,16 +166,5 @@ workflow HvpViralPipeline {
         Int   n_no_repeats                  = t_06_ViralOverallSummary.n_no_repeats
         Float mean_completeness             = t_06_ViralOverallSummary.mean_completeness
         Float mean_contamination            = t_06_ViralOverallSummary.mean_contamination
-
-        # Optional protein-annotation stage outputs
-        File? protein_annotation_combined_faa     = t_07_ProteinAnnotation.combined_proteins_faa
-        File? protein_annotation_nr_faa           = t_07_ProteinAnnotation.nr_proteins_faa
-        File? protein_annotation_foldseek_hits_tsv = t_07_ProteinAnnotation.foldseek_hits_tsv
-        File? protein_annotation_best_hits_tsv    = t_07_ProteinAnnotation.best_hits_tsv
-        File? protein_annotation_annotated_tsv    = t_07_ProteinAnnotation.annotated_hits_tsv
-        Int?  protein_annotation_num_nr           = t_07_ProteinAnnotation.num_nr_proteins
-        Int?  protein_annotation_num_hits         = t_07_ProteinAnnotation.num_foldseek_hits
-        Int?  protein_annotation_num_queries_hit  = t_07_ProteinAnnotation.num_queries_with_hits
-        Int?  protein_annotation_num_annotated    = t_07_ProteinAnnotation.num_queries_annotated
     }
 }