On branch master

6 files changed, 53 insertions, 0 deletions

diff --git a/wqflask/maintenance/dataset/calculation/__init__.py b/wqflask/maintenance/dataset/calculation/__init__.py
new file mode 100644
index 00000000..e69de29b
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/__init__.py
diff --git a/wqflask/maintenance/dataset/calculation/correlation/__init__.py b/wqflask/maintenance/dataset/calculation/correlation/__init__.py
new file mode 100644
index 00000000..e69de29b
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/correlation/__init__.py
diff --git a/wqflask/maintenance/dataset/calculation/correlation/about.txt b/wqflask/maintenance/dataset/calculation/correlation/about.txt
new file mode 100644
index 00000000..a12f8c47
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/correlation/about.txt
@@ -0,0 +1,3 @@
+BXD
+genotype, phenotype, mRNA expression
+correlation
\ No newline at end of file
diff --git a/wqflask/maintenance/dataset/calculation/correlation/conf.ini b/wqflask/maintenance/dataset/calculation/correlation/conf.ini
new file mode 100644
index 00000000..9c23bb45
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/correlation/conf.ini
@@ -0,0 +1,2 @@
+[configuration]
+genofile = /home/leiyan/gn/web/genotypes/BXD.geno
diff --git a/wqflask/maintenance/dataset/calculation/correlation/correlations.py b/wqflask/maintenance/dataset/calculation/correlation/correlations.py
new file mode 100644
index 00000000..b089e446
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/correlation/correlations.py
@@ -0,0 +1,47 @@
+# Author:               Lei Yan
+# Create Date:          2014-01-21
+# Last Update Date:     2014-01-24
+
+# import
+import sys
+import os
+import re
+import MySQLdb
+import ConfigParser
+
+def main(argv):
+
+    # load configuration from configuration file
+    config = ConfigParser.ConfigParser()
+    config.read(argv[1])
+    genofile = config.get('configuration', 'genofile')
+
+    # parse genofile
+    genotypes = []
+    file_geno = open(genofile, 'r')
+    for line in file_geno:
+        line = line.strip()
+        if line.startswith('#'):
+            continue
+        if line.startswith('@'):
+            continue
+        cells = line.split()
+        if line.startswith("Chr"):
+            strains = cells[4:]
+            continue
+        genotype = {}
+        genotype['chr'] = cells[0]
+        genotype['locus'] = cells[1]
+        genotype['cm'] = cells[2]
+        genotype['mb'] = cells[3]
+        genotype['values'] = cells[4:]
+        genotypes.append(genotype)
+    print "get %d strains:\t%s" % (len(strains), strains)
+    print "load %d genotypes" % len(genotypes)
+
+    # phenotypes
+
+# main
+if __name__ == "__main__":
+    main(sys.argv)
+    print "exit successfully"
diff --git a/wqflask/maintenance/dataset/calculation/correlation/run.sh b/wqflask/maintenance/dataset/calculation/correlation/run.sh
new file mode 100644
index 00000000..eccfa507
--- /dev/null
+++ b/wqflask/maintenance/dataset/calculation/correlation/run.sh
@@ -0,0 +1 @@
+/usr/bin/python correlations.py conf.ini
\ No newline at end of file