WIP: Checkpoint snakefile

2019_GDB/rules/MMSEQ_RULES

@@ -54,7 +54,8 @@ rule mmseq2_compare:
         mo2=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_metaspades_hybrid_rbh"),
         mo3=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades_hybrid_rbh"),
         mo4=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades_rbh"),
-        mo5=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid_rbh")
+        mo5=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid_rbh"),
+        mo6=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_metaspades_rbh")
     log: os.path.join(RESULTS_DIR, "annotation/mmseq2/compare.mmseq2.log")
     threads: config["mmseq2"]["threads"]
 #    conda: "../envs/cd-hit.yaml"
@@ -66,6 +67,7 @@ rule mmseq2_compare:
         {config[mmseqs][rbh]} {input.mm2} {input.mm3} {output.mo3} --min-seq-id 0.9 mmseq2_tmp --threads {threads} &&\
         {config[mmseqs][rbh]} {input.mm2} {input.mm4} {output.mo4} --min-seq-id 0.9 mmseq2_tmp --threads {threads} &&\
         {config[mmseqs][rbh]} {input.mm3} {input.mm4} {output.mo5} --min-seq-id 0.9 mmseq2_tmp --threads {threads} &&\
+        {config[mmseqs][rbh]} {input.mm4} {input.mm1} {output.mo6} --min-seq-id 0.9 mmseq2_tmp --threads {threads} &&\
         date) &> >(tee {log})
         """
 
@@ -79,13 +81,15 @@ rule mmseq2_m8_format:
         fo6=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades_hybrid_rbh"),
         fo7=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_db"),
         fo8=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades_rbh"),
-        fo9=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid_rbh")
+        fo9=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid_rbh"),
+        fo10=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_metaspades_rbh")
     output:
         form1=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_megahit.m8"),
         form2=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_metaspades_hybrid.m8"),
         form3=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades_hybrid.m8"),
         form4=os.path.join(RESULTS_DIR, "annotation/mmseq2/megahit_metaspades.m8"),
-        form5=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid.m8")
+        form5=os.path.join(RESULTS_DIR, "annotation/mmseq2/metaspades_metaspades_hybrid.m8"),
+        form6=os.path.join(RESULTS_DIR, "annotation/mmseq2/flye_metaspades.m8")
     log: os.path.join(RESULTS_DIR, "annotation/mmseq2/convertalis.mmseq2.log")
 #    conda: "../envs/cd-hit.yaml"
     shell:
@@ -96,6 +100,7 @@ rule mmseq2_m8_format:
         {config[mmseqs][convertalis]} {input.fo2} {input.fo3} {input.fo6} {output.form3} &&\
         {config[mmseqs][convertalis]} {input.fo2} {input.fo7} {input.fo8} {output.form4} &&\
         {config[mmseqs][convertalis]} {input.fo3} {input.fo7} {input.fo9} {output.form5} &&\
+        {config[mmseqs][convertalis]} {input.fo7} {input.fo1} {input.fo10} {output.form6} &&\
         date) &> >(tee {log})
         """
 
@@ -109,9 +114,10 @@ rule prepare_plot_files:
     shell:
         """
         (date &&\
-        scripts/prepare_plot_files.sh &&\
+        scripts/prepare_plot_files_w_metaspades.sh &&\
         date) &> >(tee {log})
         """
+# scripts/prepare_plot_files.sh &&\
 
 rule mmseq2_plots:
     input:
@@ -124,6 +130,7 @@ rule mmseq2_plots:
     shell:
         """
         (date &&\
-        Rscript scripts/mmseq_plots.R {input.plot1} {input.plot2} &&\
+        Rscript scripts/mmseq_plots_w_metaspades.R {input.plot1} {input.plot2} &&\
         date) &> >(tee {log})
         """
+# Rscript scripts/mmseq_plots.R {input.plot1} {input.plot2} &&\