WIP: Checkpoint snakefile

2019_GDB/rules/PLOT_RULES

+# For running the MMSEQ2 comparison of proteins after assemblies are run through prokka/prodigal
+
+import os
+#from tempfile import TemporaryDirectory
+
+configfile: "config/CONFIG.yaml"
+#DATA_DIR = config["data_dir"]
+RESULTS_DIR = config["results_dir"]
+#DB_DIR=config["db_dir"]
+BARCODES=config["barcodes"]
+#ASSEMBLERS=config["assemblers"]
+#MAPPERS=["bwa", "mm"]
+## SAMPLES=config["samples"]
+#SAMPLES=["flye", "megahit", "metaspades_hybrid"]
+#BINNING_SAMPLES=config["binning_samples"]
+#HYBRID_ASSEMBLER=config["hybrid_assembler"]
+SR_PREFIX="ONT3_MG_xx_Rashi_S11"
+
+
+#############################
+###### IGC COVERAGE  ########
+#############################
+rule all:
+        input: 
+            expand("{results_dir}/plots/genomecov/{sr_prefix}_vs_{lr_prefix}-x-{reference}_coverage.html", results_dir=RESULTS_DIR, sr_prefix=SR_PREFIX, lr_prefix=BARCODES, reference="igc"),
+            expand("{results_dir}/plots/annotation/diamond/lr_{lr_prefix}-sr{sr_prefix}-gene_length_ratio.html", results_dir=RESULTS_DIR, sr_prefix=SR_PREFIX, lr_prefix=BARCODES)
+
+
+rule igc_correlation_plot:
+    input:
+        sr_cov=os.path.join(RESULTS_DIR, "genomecov/sr/bwa_mem/{sr_prefix}-x-{reference}.avg_cov.txt"), 
+        lr_cov=os.path.join(RESULTS_DIR, "genomecov/lr/merged/{lr_prefix}/{lr_prefix}-x-{reference}.avg_cov.txt") 
+    output: os.path.join(RESULTS_DIR, "plots/genomecov/{sr_prefix}_vs_{lr_prefix}-x-{reference}_coverage.html")
+    #log: os.path.join(RESULTS_DIR, "plots/mapping/sr_lr_igc_coverage.log")
+    conda: "../../envs/r_ggplot2_datatable_tidyverse.yaml"
+    script:
+        "../scripts/sr_lr_igc_coverage_correlation.Rmd"
+
+rule gene_length_ratio_plot:
+    input:
+        sr_diamond=os.path.join(RESULTS_DIR, "annotation/diamond/megahit/{sr_prefix}/final.contigs.tsv"), 
+        lr_diamond=os.path.join(RESULTS_DIR, "annotation/diamond/flye/lr/merged/{lr_prefix}/assembly.tsv"),
+        hybrid_diamond=os.path.join(RESULTS_DIR, "annotation/diamond/metaspades_hybrid/lr_{lr_prefix}-sr_{sr_prefix}/contigs.tsv")
+    output: os.path.join(RESULTS_DIR, "plots/annotation/diamond/lr_{lr_prefix}-sr{sr_prefix}-gene_length_ratio.html")
+    #log: os.path.join(RESULTS_DIR, "plots/mapping/sr_lr_igc_coverage.log")
+    conda: "../../envs/r_ggplot2_datatable_tidyverse.yaml"
+    script:
+        "../scripts/gene_length_ratio.Rmd"