Merge branch 'local_blast'

Makefile

@@ -13,9 +13,9 @@ index:
 
 coverage:
 	rm -rf htmlcov .coverage
-	coverage run --source voseq voseq/manage.py test -v 2 public_interface --settings=voseq.settings.base
+	coverage run --source voseq voseq/manage.py test -v 2 public_interface blast_local --settings=voseq.settings.base
 	coverage report -m
 	coverage html
 
 test:
-	coverage run --source voseq voseq/manage.py test -v 2 public_interface --settings=voseq.settings.base
+	coverage run --source voseq voseq/manage.py test -v 2 public_interface blast_local --settings=voseq.settings.base

requirements/dev.txt

@@ -8,3 +8,5 @@ dataset==0.5.5
 django-haystack==2.3.1
 elasticsearch==1.2.0
 coveralls==0.5
+django-debug-toolbar==1.2.2
+biopython==1.65

test_db_dump.xml

@@ -305,7 +305,7 @@
             <field name="dateModification" xsi:nil="true" />
             <field name="notes" xsi:nil="true" />
             <field name="timestamp">2008-07-17 17:24:20</field>
-            <field name="id">4</field>
+            <field name="id">5</field>
             <field name="genbank">1</field>
         </row>
 	</table_data>

voseq/blast_local/admin.py

@@ -0,0 +1,3 @@
+from django.contrib import admin
+
+# Register your models here.

voseq/blast_local/models.py

@@ -0,0 +1,3 @@
+from django.db import models
+
+# Create your models here.

voseq/blast_local/tests.py

@@ -0,0 +1,57 @@
+import glob
+import os
+
+from django.core.management import call_command
+from django.test import TestCase
+from django.conf import settings
+
+from .utils import BLAST
+
+
+class BlastLocalTest(TestCase):
+    def setUp(self):
+        args = []
+        opts = {'dumpfile': 'test_db_dump.xml'}
+        cmd = 'migrate_db'
+        call_command(cmd, *args, **opts)
+
+        blast_type = 'local'
+        voucher_code = 'CP100-10'
+        gene_code = 'COI'
+        self.blast = BLAST(blast_type, voucher_code, gene_code)
+        self.seq_file = ''
+
+    '''
+    def test_have_blast_db(self):
+        result = self.blast.have_blast_db()
+        self.assertEqual(False, result)
+
+    def test_have_blast_db_true(self):
+        self.blast.create_blast_db()
+
+        result = self.blast.have_blast_db()
+        self.assertEqual(True, result)
+    '''
+
+    def test_save_seqs_to_file(self):
+        self.blast.save_seqs_to_file()
+        self.seq_file = os.path.join(settings.BASE_DIR,
+                                     '..',
+                                     'blast_local',
+                                     'db',
+                                     'COI_seqs.fas',
+                                     )
+        result = os.path.isfile(self.seq_file)
+        self.assertTrue(result)
+
+    def test_create_blast_db(self):
+        self.blast.save_seqs_to_file()
+        self.blast.create_blast_db()
+        files = glob.glob(
+            os.path.join(settings.BASE_DIR,
+                         '..',
+                         'blast_local',
+                         'db',
+                         'COI_seqs.fas.n*')
+        )
+        self.assertTrue(len(files) > 0)

voseq/blast_local/urls.py

@@ -0,0 +1,10 @@
+from django.conf.urls import patterns
+from django.conf.urls import url
+
+from . import views
+
+
+urlpatterns = patterns(
+    '',
+    url(r'^/(?P<voucher_code>.+)/(?P<gene_code>.+)/$', views.index, name='index'),
+)

voseq/blast_local/utils.py

@@ -0,0 +1,105 @@
+import os
+import re
+import subprocess
+
+from Bio.Blast.Applications import NcbiblastnCommandline
+from Bio.Seq import Seq
+from Bio import SeqIO
+from Bio.SeqRecord import SeqRecord
+
+from public_interface.models import Sequences
+
+
+class BLAST(object):
+    """
+    Class to handle duties related to local blast against sequences of one gene,
+    and full blast against all sequences in our database.
+    """
+    def __init__(self, blast_type, voucher_code, gene_code):
+        """
+        Type of blast to do: local, full, remote
+
+        :param blast_type: local, full, remote.
+        :param voucher_code:
+        :param gene_code:
+        """
+        self.blast_type = blast_type
+        self.voucher_code = voucher_code
+        self.gene_code = gene_code
+        self.mask = True
+        self.cwd = os.path.dirname(__file__)
+
+    def have_blast_db(self):
+        """
+        Finds out whether we already have a blast db with our sequences.
+
+        :return: True or False
+        """
+        pass
+
+    def is_blast_db_up_to_date(self):
+        """
+        Finds out whether our blast db contains all our sequences. In other
+        words, it finds out whether there are sequences in our postgres db with
+        time_created or time_edited more recent than our blast db files.
+
+        :return:
+        """
+        pass
+
+    def save_seqs_to_file(self):
+        """
+        Query sequences for each gene from our database and save them to local
+        disk.
+
+        Sets attribute `self.seq_file` containing necessary sequences from our
+        database.
+        """
+        if self.blast_type == 'local':
+            self.seq_file = os.path.join(self.cwd,
+                                         'db',
+                                         self.gene_code + "_seqs.fas",
+                                         )
+            queryset = Sequences.objects.all().filter(gene_code=self.gene_code)
+
+            my_records = []
+            for i in queryset:
+                id = i.code_id + '|' + i.gene_code
+                seq = self.strip_question_marks(i.sequences)
+                seq_record = SeqRecord(Seq(seq),
+                                       id=id)
+                my_records.append(seq_record)
+            SeqIO.write(my_records, self.seq_file, "fasta")
+
+    def strip_question_marks(self, seq):
+        seq = re.sub('^\?+', '', seq)
+        seq = re.sub('\?+$', '', seq)
+        return seq
+
+    def create_blast_db(self):
+        """
+        Creates a BLAST database from our sequences file in FASTA format.
+        Optionally eliminates low-complexity regions from the sequences.
+
+        :return:
+        """
+        if self.mask is True:
+            command = 'dustmasker -in ' + self.seq_file + ' -infmt fasta '
+            command += '-outfmt maskinfo_asn1_bin -out ' + self.seq_file + '_dust.asnb'
+            subprocess.check_output(command, shell=True)  # identifying low-complexity regions.
+
+            command = 'makeblastdb -in ' + self.seq_file + ' -input_type fasta -dbtype nucl '
+            command += '-mask_data ' + self.seq_file + '_dust.asnb '
+            command += '-out ' + self.seq_file + ' -title "Whole Genome without low-complexity regions"'
+            print("creating database...")
+            subprocess.check_output(command, shell=True)  # Overwriting the genome file.
+        else:
+            command = 'makeblastdb -in ' + self.seq_file + ' -input_type fasta -dbtype nucl '
+            command += '-out ' + self.seq_file + ' -title "Whole Genome unmasked"'
+            print("creating database...")
+            subprocess.check_output(command, shell=True)
+
+    def do_blast(self):
+        blastn_cline = NcbiblastnCommandline(query=self.query, db=self.db,
+                                             evalue=0.001, outfmt=5, out="opuntia.xml")
+        blastn_cline()

voseq/blast_local/views.py

@@ -0,0 +1,7 @@
+
+
+from django.shortcuts import render
+
+
+def index(request, voucher_code, gene_code):
+    return render(request, 'public_interface/base.html')

voseq/public_interface/templates/public_interface/show_voucher.html

@@ -274,7 +274,7 @@ <h3 class="panel-title"><b>Sequence Information</b></h3>
               <td><b>Lab.</b></td>
               <td><b>Accession</b></td>
               <td><b>local Blast</b></td>
-              <td><b>ncbi Blast</b></td>
+              <td><b>NCBI Blast</b></td>
             </tr>
             {% for item in sequences %}
             <tr>
@@ -283,6 +283,12 @@ <h3 class="panel-title"><b>Sequence Information</b></h3>
               <td>{{ item.ambiguous_seq_length }}</td>
               <td>{{ item.labPerson }}</td>
               <td><a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Search&db=nucleotide&term={{ item.accession }}[accn]&doptcmdl=GenBank">{{ item.accession }}</a></td>
+              <td class="text-center">
+                <a href="/blast_local/{{ voucher.code }}/{{ item.gene_code }}/"><i class="fa fa-bomb"></i></a>
+                &nbsp;&nbsp;
+                <i class="fa fa-database"></i>
+              </td>
+              <td class="text-center"><i class="fa fa-bomb"></i></td>
             </tr>
             {% endfor %}
           </table>

voseq/voseq/settings/base.py

@@ -38,6 +38,7 @@
 
     # my apps
     'public_interface',
+    'blast_local',
 )
 
 MIDDLEWARE_CLASSES = (
@@ -98,7 +99,7 @@
 
 STATIC_URL = '/static/'
 STATICFILES_DIRS = (
-    os.path.join(BASE_DIR, 'static'),
+    BASE_DIR,
 )
 
 # Get your API key from here:

voseq/voseq/settings/local.py

@@ -1,7 +1,5 @@
 import json
-import os
 
-from unipath import Path
 from django.core.exceptions import ImproperlyConfigured
 
 from .base import *
@@ -10,8 +8,7 @@
 DEBUG = True
 TEMPLATE_DEBUG = True
 
-BASE_DIR = Path(__file__).absolute().ancestor(3)
-SECRETS_FILE = os.path.join(BASE_DIR.ancestor(1), 'config.json')
+SECRETS_FILE = os.path.join(BASE_DIR, '..', '..', 'config.json')
 
 with open(SECRETS_FILE) as f:
     secrets = json.loads(f.read())
@@ -38,3 +35,5 @@ def get_secret(setting, secrets=secrets):
         'PORT': get_secret('DB_PORT'),
     }
 }
+
+INSTALLED_APPS += ('debug_toolbar',)

voseq/voseq/urls.py

@@ -3,6 +3,7 @@
 
 urlpatterns = patterns(
     '',
+    url(r'^blast_local', include('blast_local.urls', namespace='blast_local')),
     url(r'^', include('public_interface.urls', namespace='public_interace')),
     # Examples:
     # url(r'^blog/', include('blog.urls')),