From ec25e9fd2df32b547be82816b567052951e91e15 Mon Sep 17 00:00:00 2001
From: Tanushree Tunstall <tanu@tunstall.in>
Date: Sun, 16 Feb 2020 14:36:51 +0000
Subject: [PATCH] adding scripts for struct params

---
 meta_data_analysis/RD.py           | 59 ++++++++++++++++++++++++
 meta_data_analysis/dssp_df.py      | 68 ++++++++++++++++++++++++++++
 meta_data_analysis/kd.py           | 20 +++++++++
 meta_data_analysis/run_pdb_dssp.py | 72 ++++++++++++++++++++++++++++++
 meta_data_analysis/run_pdb_dssp.sh | 51 +++++++++++++++++++++
 5 files changed, 270 insertions(+)
 create mode 100755 meta_data_analysis/RD.py
 create mode 100755 meta_data_analysis/dssp_df.py
 create mode 100644 meta_data_analysis/kd.py
 create mode 100755 meta_data_analysis/run_pdb_dssp.py
 create mode 100755 meta_data_analysis/run_pdb_dssp.sh

diff --git a/meta_data_analysis/RD.py b/meta_data_analysis/RD.py
new file mode 100755
index 0000000..5bb5e6b
--- /dev/null
+++ b/meta_data_analysis/RD.py
@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Created on Thu Feb  6 12:18:24 2020
+
+@author: tanu
+"""
+#http://foldxsuite.crg.eu/faq-page#
+# after fold x downlaoded, extract and run it from 
+#https://biopython.org/DIST/docs/api/Bio.PDB.ResidueDepth%27-module.html
+#proDepth: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0007072
+#Depth server: http://cospi.iiserpune.ac.in/depth/htdocs/index.html
+# needs biopython and msms
+
+# load libraries 
+import sys, os
+import pandas as pd
+from Bio.PDB.ResidueDepth import ResidueDepth
+from Bio.PDB.PDBParser import PDBParser
+from Bio.PDB.ResidueDepth import get_surface
+
+#%%
+homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
+os.getcwd()
+os.chdir(homedir + '/git/LSHTM_analysis/meta_data_analysis/struct_params')
+os.getcwd()
+#%%
+parser = PDBParser()
+structure = parser.get_structure("3pl1", "/home/tanu/git/3pl1.pdb")
+model = structure[0]
+surface = get_surface(model)
+
+rd = ResidueDepth(model)
+print(rd['A',(' ', 152, ' ')])
+rd.keys()
+foo = rd.property_dict
+rd.property_keys
+baz = rd.property_list
+
+
+#To calculate the residue depth (average atom depth of the atoms in a residue):
+from Bio.PDB.ResidueDepth import residue_depth
+chain = model['A']
+res152 = chain[152]
+rd2 = residue_depth(res152, surface)
+
+# df from dict
+foo1 = pd.DataFrame.from_dict(baz, orient='index', columns = ['res_depth', 'surface'])
+test = pd.Series(foo, name = "test")
+
+# df from list
+foo2 = pd.DataFrame(baz, columns = ['residue', 'residue depth'])
+
+
+### iterate
+for i in range(185):
+    print(i)
+    rd3 = residue_depth(res+i, surface)
+    print(rd3)
\ No newline at end of file
diff --git a/meta_data_analysis/dssp_df.py b/meta_data_analysis/dssp_df.py
new file mode 100755
index 0000000..2791a27
--- /dev/null
+++ b/meta_data_analysis/dssp_df.py
@@ -0,0 +1,68 @@
+#!/home/tanu/anaconda3/envs/ContactMap/bin/python3
+# Read a DSSP file into a data frame and pretty-print it
+
+#https://jbloomlab.github.io/dms_tools2/dms_tools2.dssp.html
+import sys, os
+import pandas as pd
+import pprint as pp
+import dms_tools2
+import dms_tools2.dssp
+
+#%%
+# my working dir
+homedir = os.path.expanduser('~')  # spyder/python doesn't recognise tilde
+os.getcwd()
+os.chdir(homedir + '/git/LSHTM_analysis/meta_data_analysis/struct_params')
+os.getcwd()
+#%%
+# sample example
+dssp_file = "./3pl1.dssp"
+dssp_df = dms_tools2.dssp.processDSSP(dssp_file, chain='A')
+
+# outputs to console
+#returns df with ASA and RSA (base on Tien at al 2013 (theor.) values)
+#Link: https://en.wikipedia.org/wiki/Relative_accessible_surface_area
+pp.pprint(dssp_df)
+
+# write to csv
+dssp_df.to_csv('3pl1_dssp_df', header=True, index = False)
+
+#%% specify variables for input and output paths and filenames
+drug = "pyrazinamide"
+#gene = "pnca"
+
+datadir = homedir + "/git/Data"
+basedir =  datadir + "/" + drug + "/input"
+
+# input
+inpath = "/processed"
+in_filename  = "/3pl1.dssp"
+infile = basedir + inpath + in_filename
+#print(infile)
+
+# output file 
+outpath = "/output"
+outdir =   datadir + "/" + drug + outpath
+out_filename = "/3pl1_dssp_df"
+outfile =  outdir + out_filename
+print(outdir); print(outfile)
+
+if not os.path.exists(datadir):
+    print('Error!', datadir, 'does not exist. Please ensure it exists. Dir struc specified in README.md')
+    os.makedirs(datadir)
+    exit()
+
+if not os.path.exists(outdir):
+    print('Error!', outdir, 'does not exist.Please ensure it exists. Dir struc specified in README.md')
+    exit()
+    
+else:
+    print('Dir exists: Carrying on')
+# end of variable assignment for input and output files
+#%% <----- fixme
+dssp_file = infile
+dssp_df = dms_tools2.dssp.processDSSP(dssp_file, chain='A')
+
+#%%
+# write to csv
+dssp_df.to_csv(outfile, header=True, index = False)
diff --git a/meta_data_analysis/kd.py b/meta_data_analysis/kd.py
new file mode 100644
index 0000000..dabeb64
--- /dev/null
+++ b/meta_data_analysis/kd.py
@@ -0,0 +1,20 @@
+#!/usr/bin/python
+
+# hydrophobicity and SAA
+#https://biopython.org/DIST/docs/api/Bio.SeqUtils.ProtParamData-pysrc.html
+#https://jbloomlab.github.io/dms_tools2/dms_tools2.dssp.html
+import sys, os
+import pandas as pd
+import pprint as pp
+#import dms_tools2
+#import dms_tools2.dssp
+
+from Bio.SeqUtils import ProtParamData
+#%%
+homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
+os.getcwd()
+os.chdir(homedir + '/git/LSHTM_analysis/meta_data_analysis/struct_params')
+os.getcwd()
+#%%
+
+foo = ProtParamData.kd(3pl1.pdb)
diff --git a/meta_data_analysis/run_pdb_dssp.py b/meta_data_analysis/run_pdb_dssp.py
new file mode 100755
index 0000000..3a57f02
--- /dev/null
+++ b/meta_data_analysis/run_pdb_dssp.py
@@ -0,0 +1,72 @@
+#!/usr/bin/python
+# Read a PDB and output DSSP to console
+import sys, os
+from Bio.PDB import PDBParser
+from Bio.PDB.DSSP import DSSP
+import pandas as pd
+import pprint as pp
+
+#%%
+# TASK: read a pdb file and generate a dssp output file
+# FIXME: Pending output dssp hasn't been generated
+# needs dssp exe on linux
+# may be easier to run the dssp exe locally
+#%%
+# my working dir
+os.getcwd()
+homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
+os.chdir(homedir + '/git/LSHTM_analysis/meta_data_analysis')
+os.getcwd()
+#%%
+# specify variables for input and output paths and filenames
+drug = "pyrazinamide"
+#gene = "pnca"
+
+datadir = homedir + "/git/Data"
+basedir =  datadir + "/" + drug + "/input"
+
+# input
+inpath = "/original"
+
+# uncomment as necessary
+in_filename  = "/3pl1.pdb"
+
+infile = basedir + inpath + in_filename
+#print(infile)
+
+# output file 
+outpath = "/processed"
+outdir =   datadir + "/" + drug + outpath
+out_filename = "/3pl1.dssp"
+outfile =  outdir + out_filename
+#print(outdir)
+
+if not os.path.exists(datadir):
+    print('Error!', datadir, 'does not exist. Please ensure it exists. Dir struc specified in README.md')
+    os.makedirs(datadir)
+    exit()
+
+if not os.path.exists(outdir):
+    print('Error!', outdir, 'does not exist.Please ensure it exists. Dir struc specified in README.md')
+    exit()
+    
+else:
+    print('Dir exists: Carrying on')
+# end of variable assignment for input and output files
+#%%
+p = PDBParser()
+structure = p.get_structure("3pl1", infile)
+
+model = structure[0]
+dssp = DSSP(model, infile)
+#dssp = DSSP(model, infile, dssp='mkdssp') #incase you used DSSP2 exe
+pp.pprint(dssp)
+
+#DSSP data is accessed by a tuple - (chain id, residue id): RSA
+a_key = list(dssp.keys())[3]
+dssp[a_key]
+
+pp.pprint(dssp.keys())
+pp.pprint(dssp.property_dict)
+pp.pprint(dssp.property_keys)
+pp.pprint(dssp.property_list)
diff --git a/meta_data_analysis/run_pdb_dssp.sh b/meta_data_analysis/run_pdb_dssp.sh
new file mode 100755
index 0000000..d6050f4
--- /dev/null
+++ b/meta_data_analysis/run_pdb_dssp.sh
@@ -0,0 +1,51 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Created on Sun Feb 16 11:21:44 2020
+
+@author: tanu
+"""
+#!/usr/bin/bash
+# Run dssp exe
+
+#%%
+# specify variables for input and output paths and filenames
+drug="pyrazinamide"
+#gene = "pnca"
+
+datadir="~git/Data"
+basedir=${datadir}"/"${drug}
+echo${basedir}
+
+# input
+inpath="/original"
+
+# uncomment as necessary
+in_filename="/3pl1.pdb"
+
+infile=${basedir}${inpath}${in_filename}
+echo${infile}
+
+# output file 
+outpath="/processed"
+outdir=${datadir}"/"${drug}${outpath}
+out_filename="/3pl1.dssp"
+outfile=${outdir}${out_filename}
+echo${outdir}
+
+if not os.path.exists(datadir):
+    print('Error!', datadir, 'does not exist. Please ensure it exists. Dir struc specified in README.md')
+    os.makedirs(datadir)
+    exit()
+
+if not os.path.exists(outdir):
+    print('Error!', outdir, 'does not exist.Please ensure it exists. Dir struc specified in README.md')
+    exit()
+    
+else:
+    print('Dir exists: Carrying on')
+# end of variable assignment for input and output files
+#%%
+# ommand line args
+dssp -i 3pl1.pdb -o 3pl1.dssp
+dssp -i