Merge pull request #29 from PalMuc/dev

Dev
PalMuc · Oct 27, 2021 · bba92cc · bba92cc
2 parents 556c2ec + 386209e
commit bba92cc
Show file tree

Hide file tree

Showing 4 changed files with 45 additions and 53 deletions.
diff --git a/README.md b/README.md
@@ -18,7 +18,7 @@
   * [Pipeline processes](#Pipelie-processes)
   * [Manual](#Manual)
 * [Publication](#Publication)
-  * [Cite](#Cite)
+  * [Citation](#Citation)
 * [Funding](#Funding)
 * [Future work](#Future-work)
 * [Issues](#Issues)
@@ -44,9 +44,12 @@ TransPi documentation and examples can be found [here](https://palmuc.github.io/
 
 Preprint of TransPi including kmer, reads length, and reads quantities tests can be found [here](https://doi.org/10.1101/2021.02.18.431773). Also we tested the pipeline with over 45 samples from different phyla.
 
-## Cite
+TransPi has been peer-reviewed and recommended by Peer Community In Genomics
+(https://doi.org/10.24072/pci.genomics.100009)
 
-If you use TransPi please cite the preprint:
+## Citation
+
+If you use TransPi please cite the peer-reviewed publication:
 
 Rivera-Vicéns, R.E., García-Escudero, CA., Conci, N., Eitel, M., and Wörheide, G. (2021). TransPi – a comprehensive TRanscriptome ANalysiS PIpeline for de novo transcriptome assembly. bioRxiv 2021.02.18.431773; doi: https://doi.org/10.1101/2021.02.18.431773
 

diff --git a/TransPi.nf b/TransPi.nf
@@ -2873,7 +2873,7 @@ if (params.onlyAsm || params.onlyAnn || params.onlyEvi || params.all) {
                 publishDir "${workDir}/.versions", mode: "copy", overwrite: true, pattern: "*.version.txt"
 
                 input:
-                    file(assembly), file(annotation) from
+                    tuple file(assembly), file(annotation) from
 
                 output:
                     file("*.fasta") into

diff --git a/precheck_TransPi.sh b/precheck_TransPi.sh
@@ -15,16 +15,23 @@ source_c() {
         source ~/.bashrc
     fi
 }
+cleanConda () {
+    cd $mypwd
+    echo -e "\n\t -- Cleaning conda environment -- \n"
+    conda clean -a -y
+    echo -e "\n\t -- Done cleaning conda environment -- \n"
+}
 conda_only() {
     source_c
     #Check conda and environment
     check_conda=$( command -v conda )
     if [ "$check_conda" != "" ];then #&& [ "$ver" -gt "45" ];then
         echo -e "\n\t -- Conda seems to be installed in your system --\n"
-        ver=$( conda -V | awk '{print $2}' | cut -f 1,2 -d "." )
-        vern=4.8
+        ver=$( conda -V | awk '{print $2}' | cut -f 1,2 -d "." | tr -d "." )
+        vern=48
         if [ $( echo "$ver >= $vern" | bc -l ) -eq 1 ];then
-            echo -e "\n\t -- Conda is installed (v4.8 or higher). Checking environment... --\n"
+            echo -e "\n\t -- Conda is installed (v4.8 or higher) --\n"
+            #cleanConda
         fi
     else
         echo -e "\n\t -- Conda is not intalled --\n"
@@ -517,6 +524,23 @@ evi_c () {
         echo -e "\n\t -- EvidentialGene is already installed and in the PATH  -- \n"
     fi
 }
+buildsql_c () {
+    cd ${mypwd}
+    if [ -d DBs/sqlite_db/ ];then
+        cd DBs/sqlite_db/
+    else
+        mkdir -p DBs/sqlite_db/
+        cd DBs/sqlite_db/
+    fi
+}
+condaTrinotate () {
+    echo -e "\n\t -- Creating Trinotate conda environment -- \n"
+    conda create --mkdir --yes --quiet -n TPtrinotate -c conda-forge bioconda::trinotate=3.2.1=pl526_0
+    echo -e "\n\t -- Done with Trinotate conda environment -- \n"
+}
+condaTrinotateEnd () {
+    conda remove -n TPtrinotate --all -y
+}
 trisql_container () {
     if [ ! -e *.sqlite ];then
         echo -e "\n\n\t -- Custom sqlite database for Trinotate is not installed -- \n"
@@ -547,7 +571,7 @@ trisql_c () {
         condaRoot=$( conda info --json | grep "CONDA_ROOT" | cut -f 2 -d ":" | tr -d "," | tr -d " " | tr -d "\"" )
         if [ -f ${condaRoot}/etc/profile.d/conda.sh ];then
             source ${condaRoot}/etc/profile.d/conda.sh
-            conda activate TransPi
+            condaTrinotate
             check_sql=$( command -v Build_Trinotate_Boilerplate_SQLite_db.pl | wc -l )
             if [ $check_sql -eq 0 ];then
                 echo -e "\n\t -- Script \"Build_Trinotate_Boilerplate_SQLite_db.pl\" from Trinotate cannot be found -- \n"
@@ -557,6 +581,7 @@ trisql_c () {
                 Build_Trinotate_Boilerplate_SQLite_db.pl Trinotate
                 rm uniprot_sprot.dat.gz Pfam-A.hmm.gz
                 date -u >.lastrun.txt
+                condaTrinotateEnd
             fi
         fi
     elif [ -e *.sqlite ];then
@@ -565,15 +590,6 @@ trisql_c () {
         echo -e "\n\t -- Databases (PFAM,SwissProt,EggNOG,GO) last update: ${DB} --\n "
     fi
 }
-buildsql_c () {
-    cd ${mypwd}
-    if [ -d DBs/sqlite_db/ ];then
-        cd DBs/sqlite_db/
-    else
-        mkdir -p DBs/sqlite_db/
-        cd DBs/sqlite_db/
-    fi
-}
 pfam_c() {
     #Check PFAM files
     cd $mypwd
@@ -602,24 +618,6 @@ pfam_c() {
         fi
     fi
 }
-#temporary for buscoV4
-bus_env4 () {
-    echo -e "\n\t -- Creating BUSCO V4 environment --\n"
-    conda create -n busco4 -c conda-forge -c bioconda busco=4.1.4=py_0 -y
-    conda clean -a -y
-}
-bus4 () {
-    cd $mypwd
-    cpath=$( conda info --json | sed -n '/\"envs\":/,/\],/p' | grep "busco4" | tr -d "," | tr -d " " | tr -d "\"" )
-    if [ "$cpath" == "" ];then
-        echo -e "\n\t -- Cannot find the BUSCO V4 environment -- \n"
-        echo -e "\n\t -- Cleaning conda before installing BUSCO V4 environment -- \n"
-        conda clean -a -y
-        bus_env4
-    else
-        echo -e "\n\t -- BUSCO V4 environment found --\n"
-    fi
-}
 pfam_u() {
     cd $installDir
     if [ ! -d DBs/hmmerdb/ ];then
@@ -787,8 +785,8 @@ get_var () {
     echo "unpdate=\"$( if [ -f ${mypwd}/DBs/uniprot_db/.lastrun.txt ];then cat ${mypwd}/DBs/uniprot_db/.lastrun.txt;else echo "N/A";fi )\"" >>${mypwd}/.varfile.sh
     echo "pfdate=\"$( if [ -f ${mypwd}/DBs/hmmerdb/.lastrun.txt ];then cat ${mypwd}/DBs/hmmerdb/.lastrun.txt;else echo "N/A";fi )\"" >>${mypwd}/.varfile.sh
     echo "dbdate=\"$( if [ -f ${mypwd}/DBs/sqlite_db/.lastrun.txt ];then cat ${mypwd}/DBs/sqlite_db/.lastrun.txt;else echo "N/A";fi )\"" >>${mypwd}/.varfile.sh
-    echo "tenv=$( conda info --json | sed -n '/\"envs\":/,/\],/p' | grep -w "TransPi\"" | tr -d "," | tr -d " " )" >>${mypwd}/.varfile.sh
-    echo "cenv=$( conda info --json | sed -n '/\"envs\":/,/\],/p' | grep "busco4" | tr -d "," | tr -d " " )" >>${mypwd}/.varfile.sh
+    #echo "tenv=$( conda info --json | sed -n '/\"envs\":/,/\],/p' | grep -w "TransPi\"" | tr -d "," | tr -d " " )" >>${mypwd}/.varfile.sh
+    #echo "cenv=$( conda info --json | sed -n '/\"envs\":/,/\],/p' | grep "busco4" | tr -d "," | tr -d " " )" >>${mypwd}/.varfile.sh
     vpwd=$mypwd
     echo "mypwd=$mypwd" >>${vpwd}/.varfile.sh
     source .varfile.sh
@@ -802,8 +800,7 @@ get_var () {
     echo -e "\t SQL DB last update: \t $dbdate"
     echo -e "\t NEXTFLOW:\t\t $nextflow \n\n"
     cat ${confDir}/template.nextflow.config | sed -e "s|pipeInstall|pipeInstall=\"${mypwd}\"|" -e "s|busco4db|busco4db=\"${busco4db}\"|" -e "s|uniprot|uniprot=\"${uniprot}\"|" \
-        -e "s|uniname|uniname=\"${uniname}\"|" -e "s|pfloc|pfloc=\"${pfloc}\"|" -e "s|pfname|pfname=\"${pfname}\"|" -e "s|Tsql|Tsql=\"${Tsql}\"|" \
-        -e "s|myCondaInstall=\"\"|myCondaInstall=\"${tenv}\"|" -e "s|cenv=\"\"|cenv=\"${cenv}\"|" >nextflow.config
+        -e "s|uniname|uniname=\"${uniname}\"|" -e "s|pfloc|pfloc=\"${pfloc}\"|" -e "s|pfname|pfname=\"${pfname}\"|" -e "s|Tsql|Tsql=\"${Tsql}\"|" >nextflow.config
     rm .varfile.sh
 }
 get_var_user() {
@@ -826,8 +823,7 @@ get_var_user() {
     echo -e "\t PFAM files:\t\t $pfloc"
     echo -e "\t NEXTFLOW:\t\t $nextflow \n\n"
     cat ${confDir}/template.nextflow.config | sed -e "s|pipeInstall|pipeInstall=\"${mypwd}\"|" -e "s|busco4db|busco4db=\"${busco4db}\"|" -e "s|uniprot|uniprot=\"${uniprot}\"|" \
-        -e "s|uniname|uniname=\"${uniname}\"|" -e "s|pfloc|pfloc=\"${pfloc}\"|" -e "s|pfname|pfname=\"${pfname}\"|" -e "s|Tsql|Tsql=\"${Tsql}\"|" \
-        -e "s|myCondaInstall=\"\"|myCondaInstall=\"${tenv}\"|" -e "s|cenv=\"\"|cenv=\"${cenv}\"|" >nextflow.config
+        -e "s|uniname|uniname=\"${uniname}\"|" -e "s|pfloc|pfloc=\"${pfloc}\"|" -e "s|pfname|pfname=\"${pfname}\"|" -e "s|Tsql|Tsql=\"${Tsql}\"|" >nextflow.config
     rm .varfile.sh
 }
 container_pipeline_setup() {
@@ -869,7 +865,6 @@ conda_pipeline_setup() {
         buildsql_c
         trisql_c
         pfam_c
-        bus4
         echo -e "\n\t -- If no \"ERROR\" was found and all the neccesary databases are installed proceed to run TransPi -- \n"
         get_var
     fi

diff --git a/template.nextflow.config b/template.nextflow.config
@@ -75,12 +75,6 @@ params {
     // PATH to TransPi DBs installation
         pipeInstall
 
-    // PATH to conda installation from precheck. Leave blank is precheck was not used or you will use containers
-        myCondaInstall=""
-
-    //busco4 conda env
-        cenv=""
-
     // Uniprot database PATH
         uniprot
         uniname
@@ -122,7 +116,7 @@ params {
     // Skip fastp quality filter step
         skipFilter = false
     // Minimum reads quality for filtering in fastp
-        minQual="25"
+        minQual="5"
 
     // Filter rRNA
         rRNAfilter = false
@@ -188,7 +182,10 @@ params {
 
     // These options will change how the profiles work.
         // Run with conda installed by the precheck
-            myConda = false
+            //next 2 parameters are outdated
+                myConda = false
+                myCondaInstall=""
+
             condaActivate = false
 
         // TransPi container with all programs
@@ -261,9 +258,6 @@ process {
 // env Evidential Gene variable (only for nextflow)
 env.evi="${projectDir}/scripts/evigene"
 
-// Temp BUSCO V4 container for oneContainer option.
-params.v4container="ezlabgva/busco:v4.0.5_cv1"
-
 // Get PATH for cache environments
 params.localCacheDir = (params.envCacheDir ? "${params.envCacheDir}" : "${launchDir}")