From d1e8809a15ae74d0b847abb80f9f63f53078e1d6 Mon Sep 17 00:00:00 2001
From: Peter Amstutz
Date: Wed, 19 Aug 2020 15:15:57 -0400
Subject: Scaling pangenome generation

Arvados-DCO-1.1-Signed-off-by: Peter Amstutz <peter.amstutz@curii.com>
---
 .../pangenome-generate/pangenome-generate_spoa.cwl | 110 ++++++++++++---------
 1 file changed, 61 insertions(+), 49 deletions(-)

(limited to 'workflows/pangenome-generate/pangenome-generate_spoa.cwl')

diff --git a/workflows/pangenome-generate/pangenome-generate_spoa.cwl b/workflows/pangenome-generate/pangenome-generate_spoa.cwl
index 33bf64e..ed12254 100644
--- a/workflows/pangenome-generate/pangenome-generate_spoa.cwl
+++ b/workflows/pangenome-generate/pangenome-generate_spoa.cwl
@@ -19,9 +19,9 @@ outputs:
   odgiGraph:
     type: File
     outputSource: buildGraph/odgiGraph
-  odgiPNG:
-    type: File
-    outputSource: vizGraph/graph_image
+#  odgiPNG:
+#    type: File
+#    outputSource: vizGraph/graph_image
   spoaGFA:
     type: File
     outputSource: induceGraph/spoaGFA
@@ -34,16 +34,16 @@ outputs:
   mergedMetadata:
     type: File
     outputSource: dups2metadata/merged
-  indexed_paths:
-    type: File
-    outputSource: index_paths/indexed_paths
-  colinear_components:
-    type: Directory
-    outputSource: segment_components/colinear_components
+#  indexed_paths:
+#    type: File
+#    outputSource: index_paths/indexed_paths
+#  colinear_components:
+#    type: Directory
+#    outputSource: segment_components/colinear_components
 steps:
   dedup_and_sort_by_quality_and_len:
     in: {reads: seqs}
-    out: [reads_dedupped_sorted_by_quality_and_len, dups]
+    out: [sortedReadsFA, dups]
     run: sort_fasta_by_quality_and_len.cwl
   induceGraph:
     in:
@@ -54,19 +54,23 @@ steps:
     in: {inputGFA: induceGraph/spoaGFA}
     out: [odgiGraph]
     run: odgi-build-from-spoa-gfa.cwl
-  vizGraph:
-    in:
-      sparse_graph_index: buildGraph/odgiGraph
-      width:
-        default: 50000
-      height:
-        default: 500
-      path_per_row:
-        default: true
-      path_height:
-        default: 4
-    out: [graph_image]
-    run: ../tools/odgi/odgi_viz.cwl
+  # vizGraph:
+  #   in:
+  #     sparse_graph_index: buildGraph/odgiGraph
+  #     width:
+  #       default: 50000
+  #     height:
+  #       default: 500
+  #     path_per_row:
+  #       default: true
+  #     path_height:
+  #       default: 4
+  #   out: [graph_image]
+  #   requirements:
+  #     ResourceRequirement:
+  #       ramMin: $(15 * 1024)
+  #       outdirMin: 10
+  #   run: ../tools/odgi/odgi_viz.cwl
   odgi2rdf:
     in: {odgi: buildGraph/odgiGraph}
     out: [rdf]
@@ -77,29 +81,37 @@ steps:
       dups: dedup_and_sort_by_quality_and_len/dups
     out: [merged]
     run: dups2metadata.cwl
-  bin_paths:
-    run: ../tools/odgi/odgi_bin.cwl
-    in:
-      sparse_graph_index: buildGraph/odgiGraph
-      bin_width: bin_widths
-    scatter: bin_width
-    out: [ bins, pangenome_sequence ]
-  index_paths:
-    label: Create path index
-    run: ../tools/odgi/odgi_pathindex.cwl
-    in:
-      sparse_graph_index: buildGraph/odgiGraph
-    out: [ indexed_paths ]
-  segment_components:
-    label: Run component segmentation
-    run: ../tools/graph-genome-segmentation/component_segmentation.cwl
-    in:
-      bins: bin_paths/bins
-      cells_per_file: cells_per_file
-      pangenome_sequence:
-        source: bin_paths/pangenome_sequence
-        valueFrom: $(self[0])
-        # the bin_paths step is scattered over the bin_width array, but always using the same sparse_graph_index
-        # the pangenome_sequence that is extracted is exactly the same for the same sparse_graph_index
-        # regardless of bin_width, so we take the first pangenome_sequence as input for this step
-    out: [ colinear_components ]
+  # bin_paths:
+  #   requirements:
+  #     ResourceRequirement:
+  #       ramMin: 3000
+  #       outdirMin: 10
+  #   run: ../tools/odgi/odgi_bin.cwl
+  #   in:
+  #     sparse_graph_index: buildGraph/odgiGraph
+  #     bin_width: bin_widths
+  #   scatter: bin_width
+  #   out: [ bins, pangenome_sequence ]
+  # index_paths:
+  #   label: Create path index
+  #   requirements:
+  #     ResourceRequirement:
+  #       ramMin: 3000
+  #       outdirMin: 10
+  #   run: ../tools/odgi/odgi_pathindex.cwl
+  #   in:
+  #     sparse_graph_index: buildGraph/odgiGraph
+  #   out: [ indexed_paths ]
+  # segment_components:
+  #   label: Run component segmentation
+  #   run: ../tools/graph-genome-segmentation/component_segmentation.cwl
+  #   in:
+  #     bins: bin_paths/bins
+  #     cells_per_file: cells_per_file
+  #     pangenome_sequence:
+  #       source: bin_paths/pangenome_sequence
+  #       valueFrom: $(self[0])
+  #       # the bin_paths step is scattered over the bin_width array, but always using the same sparse_graph_index
+  #       # the pangenome_sequence that is extracted is exactly the same for the same sparse_graph_index
+  #       # regardless of bin_width, so we take the first pangenome_sequence as input for this step
+  #   out: [ colinear_components ]
-- 
cgit v1.2.3