fix for netmhcpan to work in snap

Author: dmnfarrell

README.md

@@ -2,11 +2,11 @@
 
 [![PyPI version shields.io](https://img.shields.io/pypi/v/epitopepredict.svg)](https://pypi.python.org/pypi/epitopepredict/)
 [![License: GPL v3](https://img.shields.io/badge/License-GPL%20v3-blue.svg)](https://www.gnu.org/licenses/gpl-3.0)
-[![Build: status](https://img.shields.io/travis/dmnfarrell/epitopepredict.svg)](https://travis-ci.org/dmnfarrell/epitopepredict)
+[![Documentation Status](https://readthedocs.org/projects/epitopepredict/badge/?version=latest)](https://epitopepredict.readthedocs.io/en/latest/?badge=latest)
 
 ### Background
 
-**epitopepredict** provides a standardized programmatic interface and command line tool for executing multiple epitope prediction methods. Currently this largely consists of interfaces to several MHC binding prediction, the results of which can then be processed and visualized in a consistent manner. The Tepitope module implements the TEPITOPEPan method is provided as a 'built in' method. The IEDB tools and netMHCIIpan and mhcflurry are also supported. All of these tools are free for academic use. This software runs on most linux systems. Users are recommended to use the snap package for convenience. This software is under active development particularly with a view to improve the command line and web tools. 
+**epitopepredict** provides a standardized programmatic interface and command line tool for executing multiple epitope prediction methods. Currently this largely consists of interfaces to several MHC binding prediction, the results of which can then be processed and visualized in a consistent manner. The Tepitope module implements the TEPITOPEPan method is provided as a 'built in' method. The IEDB tools and netMHCIIpan and mhcflurry are also supported. All of these tools are free for academic use. This software runs on most linux systems. Users are recommended to use the snap package for convenience. This software is under active development particularly with a view to improve the command line and web tools.
 
 Documentation is at http://epitopepredict.readthedocs.io
 

epitopepredict/app.py

@@ -26,8 +26,11 @@ def __init__(self, opts={}):
     def setup(self):
         """Setup main parameters"""
 
-        #if base.check_snap() == True:
-            #add_path()
+        if base.check_snap() == True:
+            print ('running in a snap')
+            cmd = base.set_netmhcpan_cmd()
+            print ('netmhcpan cmd set to:', cmd)
+
         pd.set_option('display.width', 120)
         #override base.defaults entries if provided in conf
         set_defaults(self.__dict__)
@@ -86,6 +89,9 @@ def run(self):
         preds = []
         for p in self.predictors:
             P = base.get_predictor(p)
+            if P.check_install() == False:
+                print ('%s not installed' %P.name)
+                continue
             savepath = os.path.join(self.path, p)
             if self.overwrite == True and os.path.exists(savepath):
                 shutil.rmtree(savepath)
@@ -387,10 +393,10 @@ def check_installed():
     return found
 
 def test_run():
-    """Test run for a sample file"""
+    """Test run for a sample file."""
 
     #installed = ','.join(check_installed())
-    installed = 'tepitope'
+    installed = 'tepitope,mhcflurry'
     path = os.path.dirname(os.path.abspath(__file__))
     options = config.baseoptions
     b=options['base']
@@ -403,13 +409,25 @@ def test_run():
     b['verbose'] = True
     b['cutoff_method'] = 'score'
     b['cutoffs'] = '5,500'
-    #b['compression'] = 'gzip'
-    b['overwrite'] = False
+    b['overwrite'] = True
     options = config.check_options(options)
     W = WorkFlow(options)
     st = W.setup()
     W.run()
 
+def test_binary():
+
+    cmd = base.set_netmhcpan_cmd()
+    cmd = cmd + ' -a HLA-A02:01 -l 9 P170919-test.fa'
+    print (cmd)
+    import subprocess
+    tmp = subprocess.check_output(cmd, shell=True, executable='/bin/bash')
+    import io
+    tmp = tmp.decode()
+    for i in tmp.split('\n'):
+        print (i)
+    return
+
 def main():
     "Run the application"
 
@@ -432,7 +450,7 @@ def main():
                         default=False, help="Neo-epitope pipeline")
     parser.add_option("-e", "--ensembl", dest="ensembl", action="store_true",
                         default=False, help="Get ensembl files for a release")
-    parser.add_option("-s", "--server", dest="server",
+    parser.add_option("-s", "--server", dest="server", action="store_true",
                         default=False, help="Run web app")
     parser.add_option("-x", "--port", dest="port", default=8000,
                         help="Port for web app, default 8000")
@@ -479,6 +497,7 @@ def main():
         import epitopepredict.tornado_serve
         epitopepredict.tornado_serve.main(opts.port)
     elif opts.test == True:
+        #test_binary()
         test_run()
     elif opts.version == True:
         from . import __version__

epitopepredict/base.py

@@ -406,20 +406,21 @@ def check_snap():
     """Check if inside a snap"""
 
     if 'SNAP_COMMON' in os.environ:
-        print ('running in a snap')
         return True
     return False
 
-def set_netmhcpan_cmd():
-    """Setup the netmhcpan command for using inside snap. Avoids using
-    tcsh script."""
+def set_netmhcpan_cmd(path=None):
+    """Setup the netmhcpan command to point directly to the binary. This is a
+    workaround for running inside snaps. Avoids using the tcsh script."""
 
     toolspath = os.path.join('/home', os.environ['USER'], 'tools')
     netmhcpath = os.path.join(toolspath, 'netMHCpan-4.0/Linux_x86_64')
+    if not os.path.exists(netmhcpath):
+        print ('you need to copy the netmhcpan files to %s' %toolspath)
     os.environ['NETMHCpan']=netmhcpath
-    os.environ['TMPDIR']='/tmp'
+    os.environ['TMPDIR']= '/tmp'
     cmd = os.path.join(netmhcpath, 'bin/netMHCpan')
-    print ('netmhcpan cmd set to:', cmd)
+    #print (os.environ)
     return cmd
 
 class DataFrameIterator:
@@ -474,6 +475,9 @@ def __repr__(self):
             n = len(self.data.name.unique())
             return '%s predictor with results in %s sequences' %(self.name, n)
 
+    def check_install(self):
+        return True
+
     def supported_lengths(self):
         """Return supported peptide lengths"""
         return [9,11]
@@ -1204,7 +1208,7 @@ def __init__(self, data=None, scoring='affinity'):
         self.qf = self.get_quantile_data()
         self.basecmd = 'netMHCpan'
         #base command needs to be to the binary directly if running snap
-        if check_snap is True:
+        if check_snap() is True:
             self.basecmd = set_netmhcpan_cmd()
 
     def read_result(self, temp):
@@ -1291,13 +1295,13 @@ def convert_allele_name(self, a):
         return a
 
     def check_install(self):
-        cmd = 'netMHCpan'
+
+        cmd = '%s -h' %self.basecmd
         try:
             temp = subprocess.check_output(cmd, shell=True)
             print('netMHCpan appears to be installed')
             return 1
         except:
-            print ('netMHCpan not found')
             return 0
 
     @classmethod
@@ -1821,7 +1825,7 @@ def __init__(self, data=None, **kwargs):
         self.operator = '<'
         self.scorekey = 'score'
         self.rankascending = 1
-        if self.check_install == False:
+        if self.check_install() == False:
             return
         self._check_models()
         from mhcflurry import Class1AffinityPredictor
@@ -1870,10 +1874,10 @@ def get_alleles(self):
 
     def check_install(self):
         try:
-            import mhcflurry
+            from mhcflurry import Class1AffinityPredictor
             return True
         except:
-            print ('use pip install mhcflurry to install.')
+            #print ('use pip install mhcflurry to install.')
             return False
 
     def _check_models(self):

epitopepredict/description.txt

@@ -1,34 +1,37 @@
 Background
 ----------
 
-epitopepredict provides a standardized programmatic interface and command line
-tool for executing several epitope prediction methods. Currently this largely
-consists of interfaces to several MHC binding prediction, the results of which
- can then be processed and visualized in a consistent manner. Many MHC binding
- predictors have been developed but usually not in an open source manner. The
- Tepitope module implements the TEPITOPEPan method is provided as a 'built in'
- method. netMHCIIpan must be downloaded separately from the website and
- installed on your system. The process is quite simple. The same applies for
- the IEDB tools. Both of these tools are free for academic use. It is hoped
- that other epitope predictors will be integrated.
+epitopepredict provides a standardized programmatic interface and command line tool for
+executing multiple epitope prediction methods. Currently this largely consists of interfaces
+to several MHC binding prediction, the results of which can then be processed and visualized
+in a consistent manner. The Tepitope module implements the TEPITOPEPan method is provided as
+a 'built in' method. The IEDB tools and netMHCIIpan and mhcflurry are also supported.
+All of these tools are free for academic use. This software runs on most linux systems.
+Users are recommended to use the snap package for convenience. This software is under
+active development particularly with a view to improve the command line and web tools.
+
+Documentation is at http://epitopepredict.readthedocs.io
 
 Supported methods
 -----------------
 
 TEPITOPEPan
-NetMHCIIpan
+NetMHCpan
 IEDB MHC tools
-IEDB BCell tools
-MHCFlurry
+MHCflurry
 
 Dependencies
 ------------
 
 pandas
 biopython
+tornado
+bokeh
+mhcflurry (optional)
 
 Installation
 ------------
+
 pip install epitopepredict
 
 Links

epitopepredict/neo.py

@@ -130,8 +130,8 @@ def run(self):
                 else:
                     alleles = self.mhc2_alleles
                     length = self.mhc2_length
-                seqs = get_mutant_sequences(effects=effects, length=length, verbose=self.verbose)
 
+                seqs = get_mutant_sequences(effects=effects, length=length, verbose=self.verbose)
                 res = predict_variants(seqs, alleles=alleles, predictor=predictor,
                                        verbose=self.verbose, cpus=self.cpus)
                 res['label'] = f
@@ -470,9 +470,8 @@ def predict_variants(df, predictor='netmhcpan', alleles=[],
     df['length'] = df.peptide.str.len()
 
     P = base.get_predictor(predictor, scoring='ligand')
-    if verbose == True:
-        print (P)
-        print ('predicting mhc binding for %s peptides with %s' %(len(df), P.name))
+    print (P)
+    print ('predicting mhc binding for %s peptides with %s' %(len(df), P.name))
 
     peps = list(df.peptide)
     res = P.predict_peptides(peps, alleles=alleles, cpus=cpus,

setup.py

@@ -20,7 +20,7 @@
                   'description.txt']
                  },
     install_requires=['numpy>=1.10',
-                      'pandas>=0.20',
+                      'pandas>=0.22',
                       'biopython>=1.5',
                       'bokeh==0.12.14',
                       'wtforms>=2.1',

snap/snapcraft.yaml

@@ -25,4 +25,4 @@ parts:
     python-packages:
       [mhcflurry,gtfparse==1.2,pyensembl==1.7.3,varcode==0.8.0]
     stage-packages:
-      [tcsh,gawk,ncbi-blast+]
+      [gawk,ncbi-blast+]