From 36521d8e99475e079b185c7924258d7a74a28130 Mon Sep 17 00:00:00 2001
From: Jay Hesselberth <jay.hesselberth@gmail.com>
Date: Fri, 10 Jan 2025 12:03:10 -0700
Subject: [PATCH] Process files for squigualiser

---
 cluster/config.yaml             |  3 +++
 environment.yml                 |  8 +++++---
 workflow/Snakefile              |  1 +
 workflow/rules/common.smk       |  9 +++++----
 workflow/rules/squigualiser.smk | 33 +++++++++++++++++++++++++++++++++
 5 files changed, 47 insertions(+), 7 deletions(-)
 create mode 100644 workflow/rules/squigualiser.smk

diff --git a/cluster/config.yaml b/cluster/config.yaml
index ace6d5f..4cb91f6 100644
--- a/cluster/config.yaml
+++ b/cluster/config.yaml
@@ -32,6 +32,9 @@ set-resources:
   - cca_classify:queue="gpu"
   - cca_classify:gpu_opts="-gpu num=1:j_exclusive=yes"
   - cca_classify:ngpu=1
+  - resquiggle:queue="gpu"
+  - resquiggle:gpu_opts="-gpu num=1:j_exclusive=yes"
+  - resquiggle:ngpu=1
 
 
 printshellcmds: True
diff --git a/environment.yml b/environment.yml
index 708ef0b..20f2087 100644
--- a/environment.yml
+++ b/environment.yml
@@ -1,4 +1,4 @@
-name: aatrnaseqpipe
+name: aa-trna-seq-pipeline
 
 channels:
   - conda-forge
@@ -17,10 +17,12 @@ dependencies:
   - gitpython
   - pip
   - dorado >= 0.7.2
+  - f5c
+  - squigualiser
   - pip:
     - pod5
     - ont-remora >= 3.2
     - snakemake-executor-plugin-lsf
     - snakemake-executor-plugin-slurm
-    - snakemake-executor-plugin-cluster-generic
-    - GPUtil
+    - snakemake-executor-plugin-cluster-generic 
+    - blue-crab
diff --git a/workflow/Snakefile b/workflow/Snakefile
index fbc6dc2..7aa409c 100644
--- a/workflow/Snakefile
+++ b/workflow/Snakefile
@@ -14,6 +14,7 @@ PIPELINE_DIR = os.path.dirname(SNAKEFILE_DIR)
 
 include: "rules/common.smk"
 include: "rules/aatrnaseq.smk"
+include: "rules/squigualiser.smk"
 
 
 report_metadata()
diff --git a/workflow/rules/common.smk b/workflow/rules/common.smk
index a5a12c5..ceb98dc 100644
--- a/workflow/rules/common.smk
+++ b/workflow/rules/common.smk
@@ -127,6 +127,11 @@ def pipeline_outputs():
         values=["cpm", "counts"],
     )
 
+    outs += expand(
+        os.path.join(outdir, "squigualizer", "{sample}", "{sample}.tsv"),
+        sample=samples.keys(),
+    )
+
     if (
         "remora_kmer_table" in config
         and config["remora_kmer_table"] != ""
@@ -137,10 +142,6 @@ def pipeline_outputs():
             sample=samples.keys(),
         )
 
-    # if "trna_table" in config and config["trna_table"] != "" and config["trna_table"] is not None:
-    #    outs += expand(os.path.join(outdir, "tables", "{sample}", "{sample}.charging_status.tsv"),
-    #        sample = samples.keys())
-
     return outs
 
 
diff --git a/workflow/rules/squigualiser.smk b/workflow/rules/squigualiser.smk
new file mode 100644
index 0000000..c84092c
--- /dev/null
+++ b/workflow/rules/squigualiser.smk
@@ -0,0 +1,33 @@
+
+rule blue_crab:
+    input:
+        rules.merge_pods.output,
+    output:
+        os.path.join(outdir, "squigualizer", "{sample}", "{sample}.blow5"),
+    log:
+        os.path.join(outdir, "logs", "blue_crab", "{sample}"),
+    shell:
+        """
+        blue-crab p2s {input} -o {output}
+        """
+
+
+rule resquiggle:
+    input:
+        fastq=rules.ubam_to_fq.output,
+        signal=rules.blue_crab.output,
+    output:
+        os.path.join(outdir, "squigualizer", "{sample}", "{sample}.tsv"),
+    log:
+        os.path.join(outdir, "logs", "resquiggle", "{sample}"),
+    shell:
+        """
+        f5c_x86_64_linux_cuda \
+            resquiggle \
+            # XXX -B depends on memory available
+            -B 10 \
+            --rna \
+            -c {input.fastq} \
+            {input.signal} \
+            -o {output} 
+        """