ul97 and hard fix some dependencies (#23)

* ul97 and hard fix some dependencies

* integrate pfam script

* disorder back

.gitignore

@@ -6,3 +6,4 @@ t/
 query/
 lib/
 test/
+help.sh

Dockerfile

@@ -2,64 +2,65 @@ FROM docker.io/ubuntu:22.04
 
 
 # setup
-ENV ENV TZ=Europe/London
-RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
+ENV ENV TZ=Europe/London ; ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone ; hwclock --hctosys
 RUN apt-get update
+ARG DEBIAN_FRONTEND=noninteractive
 
 
 ### general dependencies
-RUN apt install -y default-jdk
-RUN apt install -y wget
-RUN apt install -y git
-RUN apt install -y cmake
-RUN apt install -y mrc
-RUN apt install -y libboost-all-dev 
-RUN apt install -y libpcre2-8-0 libpcre2-dev liblzma-dev libbz2-dev
-RUN apt install -y python3.10
-RUN apt install -y python3-pip
-RUN apt install -y python3.10-venv
-RUN apt install -y --no-install-recommends r-base-core
-RUN apt install -y mafft
-RUN apt install -y python2
-
-
-### data generating tools / dependencies
-# psiblast
-RUN apt install -y ncbi-blast+
-# req of p2rank
-RUN apt install -y python3-pymol
-# netsurfp3 - ref: https://dtu.biolib.com/NetSurfP-3/
-RUN python3 -m pip install -U pybiolib
+RUN apt install -y default-jdk wget git cmake mrc libboost-all-dev libpcre2-dev liblzma-dev libbz2-dev nano curl libcurl4-openssl-dev libeigen3-dev
+RUN apt install -y python3.10 python3-pip python3.10-venv python2 python3-pymol
+
+
+### scientific informatics
+RUN apt install -y ncbi-blast+ mafft hmmer
+
 
 # R packages
-RUN apt install -y r-cran-stringr r-cran-reshape2 r-cran-ggpubr r-cran-tidyr \
-    r-cran-readr r-cran-ape r-cran-devtools r-cran-biocmanager 
-RUN R CMD javareconf
-RUN apt install -y curl libcurl4-openssl-dev libeigen3-dev
+RUN apt install -y --no-install-recommends r-base-core r-cran-stringr r-cran-reshape2 r-cran-ggpubr r-cran-tidyr \
+    r-cran-readr r-cran-ape r-cran-devtools r-cran-biocmanager ; R CMD javareconf
 COPY ./scripts/ /scripts/
 COPY ./lib/ /mflibs/
 RUN Rscript /scripts/install_r_packages.R
-RUN /scripts/install_stuff.sh
 
+RUN mkdir /tools && \
+    git clone https://github.com/PDB-REDO/libcifpp && \
+    cd libcifpp && \
+    git checkout 288b2bb72093054f9b66604fd7dca4a3a6ea0a27 && \
+    mkdir build && \
+    cd build && \
+    cmake ..  && \
+    cmake --build . --config Release && \
+    cmake --install .
+RUN cd /tools && \
+    git clone https://github.com/mhekkel/libmcfp.git && \
+    cd libmcfp && \
+    git reset --hard 4aa95505ded43e663fd9dae61c49b08fdc6cce0c && \
+    mkdir build && \
+    cd build && \
+    cmake .. && \
+    cmake --build . && \
+    cmake --install .
+
+RUN cd /tools && \
+    git clone https://github.com/PDB-REDO/dssp.git && \
+    cd dssp && \
+    git checkout db629fb2282a5d9c58048d6ca833027e5a214cf3 && \
+    cmake -S . -B build && \
+    cmake --build build && \
+    cmake --install build
 
 
 RUN bash /scripts/pdb2ProtLigSite.sh -s
 RUN bash /scripts/msa2coupling.sh -s
+RUN bash /scripts/Seq2Disorder.sh -s
 #RUN bash /scripts/Seq2ProtLangRep.sh -s
-#RUN bash /scripts/Seq2Disorder.sh -s
 #RUN bash /scripts/Seq2SecStruc.sh -s
 
-RUN python3 -m pip install biopython numpy pandas pybiolib
-RUN apt install -y nano
-RUN apt install -y hmmer
-
-# dev
-COPY Seq2PfamResidues.sh /scripts
-RUN bash /scripts/Seq2PfamResidues.sh -s
-
-
 COPY mf.R /scripts
 
+RUN python3 -m pip install numpy pandas Bio
+
 
 #ENTRYPOINT ["/bin/bash", "-c"]
 CMD ["/bin/bash"]

mf.R

@@ -5,7 +5,7 @@
 # -------------------- Setup
 ### runtime vars
 args = commandArgs(trailingOnly=TRUE)
-infasta = as.character(args[1]) #"/query/HCMV_UL54.fasta"
+infasta = as.character(args[1]) #"/query/HCMV_UL97.fasta"
 blast_db_name = as.character(args[2]) # "uniref50.fasta"
 threads = as.numeric(args[3]) # 32
 v_eval = as.character(args[4]) # 1e-7 # psiblast e value
@@ -70,6 +70,10 @@ for(i in 1:nrow(t)){
 }
 writeLines(out_fasta, temp_blast)
 
+# deduplicate fasta
+system( paste0("awk '/^>/{f=!d[$1];d[$1]=1}f' ", temp_blast, " > /tmp/temp.fa") )
+system( paste0("cp /tmp/temp.fa ", temp_blast) )
+
 # align outputted fasta
 command = paste0("mafft --add  ", temp_blast," --keeplength --thread ",  threads ," ",infasta, " > ", temp_blast_msa," 2>",tdir,"/err.txt" )
 system(command)
@@ -399,7 +403,7 @@ df = cbind(df,physdat)
 
 
 # ------------------------------------------------------------ Structural features
-if(1 == 2){
+if(1 == 1){
   struc = list()
   # --------------------  from sequence
   ### disorder