outstanding changes for working version

Author: ctastad

workflow/Snakefile

@@ -31,9 +31,7 @@ rule to_spatialdata:
         paths.sdata_zgroup if paths.data_path else [],
     conda:
         "sopa2"
-    threads: 2
-    resources:
-        lsf_extra="-gpu num=1"
+    threads: 4
     params:
         reader = args['read'].as_cli(),
         data_path = paths.data_path,
@@ -50,7 +48,7 @@ rule tissue_segmentation:
         touch(paths.smk_tissue_segmentation),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         tissue_segmentation = args["segmentation"]["tissue"].as_cli(),
         sdata_path = paths.sdata_path,
@@ -71,7 +69,7 @@ checkpoint patchify_image:
         sdata_path = paths.sdata_path,
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     shell:
         """
         sopa patchify image {params.sdata_path} {params.patchify_image}
@@ -90,7 +88,7 @@ checkpoint patchify_transcripts:
         sdata_path = paths.sdata_path,
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     shell:
         """
         sopa patchify transcripts {params.sdata_path} {params.patchify_transcripts}
@@ -104,7 +102,6 @@ rule patch_segmentation_cellpose:
         paths.smk_cellpose_temp_dir / "{index}.parquet",
     conda:
         "sopa2"
-    threads: 2
     resources:
         lsf_extra="-gpu num=1"
     params:
@@ -127,6 +124,8 @@ rule patch_segmentation_baysor:
         sdata_path = paths.sdata_path,
     shell:
         """
+        ml gcc/11.2.0
+
         export JULIA_NUM_THREADS={resources.cpus_per_task} # parallelize within each patch for Baysor >= v0.7
 
 
@@ -167,7 +166,7 @@ rule resolve_cellpose:
         touch(paths.smk_cellpose_boundaries),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         sdata_path = paths.sdata_path,
     shell:
@@ -183,7 +182,7 @@ rule resolve_baysor:
         touch(paths.smk_table),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         resolve = args.resolve_transcripts(),
         sdata_path = paths.sdata_path,
@@ -203,7 +202,7 @@ rule resolve_comseg:
         touch(paths.smk_table),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         resolve = args.resolve_transcripts(),
         sdata_path = paths.sdata_path,
@@ -232,12 +231,18 @@ rule aggregate:
         touch(paths.smk_aggregation),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         aggregate = args["aggregate"].as_cli(),
         sdata_path = paths.sdata_path,
+        gene_exclude_pattern = config.get("gene_exclude_pattern", "negcontrol.*|blank.*|antisense.*|unassigned.*|deprecated.*|intergenic.*"),
     shell:
         """
+        python -c "
+import sopa
+pattern = '{params.gene_exclude_pattern}'
+sopa.settings.gene_exclude_pattern = None if pattern == 'None' else pattern
+"
         sopa aggregate {params.sdata_path} {params.aggregate}
         """
 
@@ -248,7 +253,7 @@ rule annotate:
         directory(paths.annotations),
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     resources:
         lsf_extra="-gpu num=1"
     params:
@@ -289,7 +294,7 @@ rule image_write:
         paths.explorer_image,
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         explorer = args["explorer"].as_cli(keys=['lazy', 'ram_threshold_gb', 'pixel_size', 'pixelsize']),
         sdata_path = paths.sdata_path,
@@ -312,7 +317,7 @@ rule report:
         report = paths.report,
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     shell:
         """
         sopa report {params.sdata_path} {params.report}
@@ -328,7 +333,7 @@ rule explorer:
         paths.explorer_experiment,
     conda:
         "sopa2"
-    threads: 2
+    threads: 4
     params:
         explorer = args["explorer"].as_cli(),
         sdata_path = paths.sdata_path,

workflow/profile/lsf/config.yaml

@@ -4,7 +4,8 @@ default-resources:
     gpu: 0
     lsf_project: acc_untreatedIBD
     lsf_queue: premium
-    runtime: 60
+    runtime: 120
+    cpus_per_task: 4
 
 set-resources:
     to_spatialdata:
@@ -20,14 +21,14 @@ set-resources:
     annotate:
         lsf_queue: gpu
         mem_mb: 64_000
-        runtime: 180
+        runtime: 240
     novae:
         lsf_queue: gpu
         mem_mb: 64_000
         runtime: 45
+        cpus_per_task: 16
     patch_segmentation_baysor:
         mem_mb: 32_000
-        cpus_per_task: 2
     resolve_baysor:
         mem_mb: 128_000
     aggregate: