🚧 Use augur subsample

phylogenetic/defaults/config_zika.yaml

@@ -2,11 +2,10 @@ strain_id_field: "accession"
 display_strain_field: "strain"
 
 filter:
-  group_by: "country year month"
-  sequences_per_group: 40
-  min_date: 2012
   min_length: 5385
 
+subsampling: "defaults/subsampling.yaml"
+
 refine:
   coalescent: "opt"
   date_inference: "marginal"

phylogenetic/defaults/subsampling.yaml

@@ -0,0 +1,8 @@
+samples:
+  all:
+    group_by:
+    - country
+    - year
+    - month
+    sequences_per_group: 40
+    min_date: 2012

phylogenetic/rules/prepare_sequences.smk

@@ -52,21 +52,17 @@ rule decompress:
 rule filter:
     """
     Filtering to
-      - {params.sequences_per_group} sequence(s) per {params.group_by!s}
-      - from {params.min_date} onwards
-      - excluding strains in {input.exclude}
+      - exclude strains in {input.exclude}
       - minimum genome length of {params.min_length} (50% of Zika virus genome)
     """
     input:
         sequences = "data/sequences_all.fasta",
         metadata = "data/metadata_all.tsv",
         exclude = "defaults/dropped_strains.txt",
     output:
+        metadata = "results/filtered.tsv",
         sequences = "results/filtered.fasta"
     params:
-        group_by = config["filter"]["group_by"],
-        sequences_per_group = config["filter"]["sequences_per_group"],
-        min_date = config["filter"]["min_date"],
         min_length = config["filter"]["min_length"],
         strain_id = config.get("strain_id_field", "strain"),
     shell:
@@ -77,19 +73,41 @@ rule filter:
             --metadata-id-columns {params.strain_id} \
             --exclude {input.exclude} \
             --output {output.sequences} \
-            --group-by {params.group_by} \
-            --sequences-per-group {params.sequences_per_group} \
-            --min-date {params.min_date} \
+            --output-metadata {output.metadata} \
             --min-length {params.min_length}
         """
 
+rule subsample:
+    """
+    Subsampling with config defined in {params.config}.
+    """
+    input:
+        metadata = "results/filtered.tsv",
+        sequences = "results/filtered.fasta",
+    output:
+        metadata = "results/subsampled.tsv",
+        sequences = "results/subsampled.fasta",
+    params:
+        config = config["subsampling"],
+        strain_id = config.get("strain_id_field", "strain"),
+    shell:
+        """
+        augur subsample \
+            --sequences {input.sequences} \
+            --metadata {input.metadata} \
+            --config {params.config} \
+            --metadata-id-columns {params.strain_id} \
+            --output-metadata {output.metadata} \
+            --output-sequences {output.sequences}
+        """
+
 rule align:
     """
     Aligning sequences to {input.reference}
       - filling gaps with N
     """
     input:
-        sequences = "results/filtered.fasta",
+        sequences = "results/subsampled.fasta",
         reference = "defaults/zika_reference.gb"
     output:
         alignment = "results/aligned.fasta"