remove deprecated fields used for POC

web/metagenedb/apps/catalog/migrations/0001_initial.py

-# Generated by Django 2.2.1 on 2019-06-07 08:47
-
-from django.db import migrations, models
-
-
-class Migration(migrations.Migration):
-
-    initial = True
-
-    dependencies = [
-    ]
-
-    operations = [
-        migrations.CreateModel(
-            name='Gene',
-            fields=[
-                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
-                ('gene_id', models.CharField(db_index=True, max_length=100, unique=True)),
-                ('gene_length', models.IntegerField()),
-                ('taxonomic_genus', models.CharField(max_length=100)),
-                ('taxonomic_phylum', models.CharField(max_length=100)),
-            ],
-        ),
-    ]

web/metagenedb/apps/catalog/models.py

@@ -4,8 +4,6 @@ from django.db import models
 class Gene(models.Model):
     gene_id = models.CharField(max_length=100, unique=True, db_index=True)
     gene_length = models.IntegerField()
-    taxonomic_genus = models.CharField(max_length=100)
-    taxonomic_phylum = models.CharField(max_length=100)
 
     def __str__(self):
         return self.gene_id

web/metagenedb/apps/catalog/serializers.py

@@ -6,4 +6,4 @@ class GeneSerializer(serializers.ModelSerializer):
 
     class Meta:
         model = Gene
-        fields = ('gene_id', 'gene_length', 'taxonomic_genus', 'taxonomic_phylum')
+        fields = ('gene_id', 'gene_length')

web/scripts/import_igc_data.py

@@ -18,11 +18,27 @@ _LOGGER = logging.getLogger(__name__)
 
 
 def create_gene(raw_line):
+    """
+    IGC annotation columns:
+        0: Gene ID	                            Unique ID
+        1: Gene Name                 	        Unique name
+        2: Gene Length	                        Length of nucleotide sequence
+        3: Gene Completeness Status	            Stating a gene is complete or partial according to the gene predictor
+        4: Cohort Origin	                    Stating the cohort contributing the representative gene
+        5: Taxonomic Annotation(Phylum Level)	Annotated phylum for a gene
+        6: Taxonomic Annotation(Genus Level)	Annotated genus for a gene
+        7: KEGG Annotation	                    Annotated KO(s) for a gene
+        8: eggNOG Annotation	                Annotated eggNOG(s) for a gene
+        9: Sample Occurence Frequency	        Occurrence frequency in samples based on gene profile
+        10:Individual Occurence Frequency	    Occurrence frequency in individuals based on gene profile
+        11: KEGG Functional Categories	        KEGG functional category(ies) of the annotated KO(s)
+        12: eggNOG Functional Categories	    eggNOG functional category(ies) of the annotated eggNOG(s)
+        13: Cohort Assembled	                Stating the metagenomic sequencing cohort(s) contributing the
+                                                representative gene or a redundant gene belonging to it
+    """
     gene_info = raw_line.rstrip().split('\t')
     gene = Gene(gene_id=gene_info[1],
-                gene_length=gene_info[2],
-                taxonomic_genus=gene_info[6],
-                taxonomic_phylum=gene_info[5])
+                gene_length=gene_info[2])
     return gene