script 5 in progress, cat

script/4_parse_PSIblast.R

@@ -101,6 +101,7 @@ system(cmd)
 library(ape)
 
 aln<-read.dna(fasta, format="fasta")
+aln<-clustal(aln)
 
 names(listid_short)<-gsub(names(listid_short), pattern=" ", replacement="_")
 

script/4_parse_PSIblast.sh

@@ -68,7 +68,6 @@ nrow=`cat $subblast | wc -l`
     echo "sequences already retrieved"
     else
     echo "retrieve fasta"
-    #Rscript --vanilla $rscript $subblast $EVAL $PERCID $PERCOVERLAP $NPERTAX $FASTA> $FASTA_REP/$Gene/paste_blast_test.log
     Rscript --vanilla $rscript $subblast $EVAL $PERCID $PERCOVERLAP $NPERTAX $FASTA $BLASTDBNUC $TAXO > $FASTA_REP/$Gene/paste_blast_test.log
     fi ;
 

script/5_cat_aln_phy.R

+#!/usr/bin/env Rscript
+args = commandArgs(trailingOnly=TRUE)
+
+# test if there is at least one argument: if not, return an error
+if (length(args)!=1) {
+  stop("1 arguments must be supplied.\n", call.=FALSE)
+}
+
+
+#args<-c("/home/mcariou/2021_legio/fasta/78Lp", "~/2020_Attaiech/prot_db/Transdecoder", "~/2021_legio/phylolegio/doc/tabAss.txt")
+#args<-c("/home/mcariou/2021_legio/fasta/78Lp")
+
+
+subrep<-list.files(args[1])
+subrep<-subrep[grep(x=subrep, pattern="Q5")]
+
+
+#### Change names in the fasta
+library(ape)
+
+
+cat<-list()
+
+
+#### fuction
+
+file<-list.files(paste0(args[1], "/", subrep[1]))
+file<-file[grep(x=file, pattern="aln")]
+aln<-read.dna(paste0(args[1], "/",subrep[1], "/", file), format="fasta", as.character=TRUE)
+
+tmp<-aln
+names(tmp)<-sapply(labels(aln), function(x) strsplit(x, split=".", fixed=TRUE)[[1]][1])
+
+for (seq in names(tmp)){
+    print(seq)
+    cat<-c(cat, list(seq=tmp[seq]))
+    names(cat)<-c(names(cat)[-length(cat)], seq)
+}
+
+
+
+#write.dna(aln, file=fasta_aln, format="fasta", nbcol = 6, colsep = "", colw = 10)
+
+
+
+
+

script/5_cat_aln_phy.sh

+#!/bin/bash
+
+echo "USAGE: ./5_aln_phy.sh \$1=fasta_path"
+echo "--------------------------------------------------"
+
+
+##################################################################################################################
+### En local 
+#./5_aln_phy.sh ~/Documents/CIRI_BIBS_projects/2021_04_Doublet/pipeline/fasta/ lp0952_ortho
+
+### PSMN
+#./5_cat_aln_phy.sh ~/2021_legio/fasta/78Lp/
+##################################################################################################################
+
+# variable
+DATA=$1
+FASTA=$1/concat.fasta
+PHYLIP=$1/concat.phylip
+ 
+echo $0
+rscript=`ls $0 | sed 's/.sh/.R/g'`
+echo $rscript
+
+
+
+if [[ -s $FASTA ]] ; then
+    echo "sequences already concatenated"
+else
+    echo "concatenate fasta"
+    Rscript --vanilla $rscript $DATA > $DATA/cat.log
+fi ;
+
+
+if [ -e $FASTA ] ; then
+  echo $FASTA" exists"
+
+  #prank -d=$FASTA +F -o=$ALN
+  echo $ALN
+  ## Convert to phylipX
+  trimal -in $FASTA -out $PHYLIP -phylip
+
+  ## phylogeny
+  phyml -i $PHYLIP -d nt -m HKY85 -a e -c 4 -s NNI -b -1
+
+else
+  echo $FASTA "do not exists. incorrect input"
+fi ;  
+
+
+
+
+
+# fin