Mercurial > repos > melissacline > ucsc_cancer_utilities

changeset 54:59dbe857f5d4

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
introduce normal_CNV parameter
author jingchunzhu
date Thu, 17 Sep 2015 22:03:04 -0700
parents 7de3db823f90
children 1093078e7976
files seg2matrix/CGData/CsegToMatrix.cc seg2matrix/CGData/SegToMatrix.py seg2matrix/mapSegToGeneMatrix.py seg2matrix/segToMatrixGalaxy.py segToGeneMatrix.xml segToMatrix.xml
diffstat 6 files changed, 22 insertions(+), 18 deletions(-) [+]
line wrap: on
line diff
--- a/seg2matrix/CGData/CsegToMatrix.cc	Thu Sep 17 15:38:53 2015 -0700
+++ b/seg2matrix/CGData/CsegToMatrix.cc	Thu Sep 17 22:03:04 2015 -0700
@@ -77,7 +77,7 @@
 
 
 
-void print_matrix(segmap *data, set<string> *targetSet, void (*print)(const char *)) {
+void print_matrix(segmap *data, set<string> *targetSet, void (*print)(const char *), char* NORMAL_CNV) {
 	
 	//create a break map
 	breakmap breaks;
@@ -176,7 +176,7 @@
 				float val = gm[i][j];
 				print("\t");
 				if ( val == MISSING_VAL )
-					print("NA");
+					print(NORMAL_CNV);
 				else {							
 					char str[20]  = "";
 					sprintf(str, "%f", val);
--- a/seg2matrix/CGData/SegToMatrix.py	Thu Sep 17 15:38:53 2015 -0700
+++ b/seg2matrix/CGData/SegToMatrix.py	Thu Sep 17 22:03:04 2015 -0700
@@ -10,8 +10,7 @@
 if os.path.exists(libFile):
     segLib = ctypes.cdll.LoadLibrary(libFile)
 
-
-def seg_to_matrix(seg_handle, out_handle):
+def seg_to_matrix(seg_handle, out_handle, NORMAL_CNV):
     """
     Turn a segment file into a segmented matrix.
     
@@ -29,4 +28,4 @@
         return 0
 
     printbackTYPE = ctypes.CFUNCTYPE(ctypes.c_int, ctypes.c_char_p)
-    segLib.print_matrix(s, t, printbackTYPE(printback))
+    segLib.print_matrix(s, t, printbackTYPE(printback), NORMAL_CNV)
--- a/seg2matrix/mapSegToGeneMatrix.py	Thu Sep 17 15:38:53 2015 -0700
+++ b/seg2matrix/mapSegToGeneMatrix.py	Thu Sep 17 22:03:04 2015 -0700
@@ -7,12 +7,14 @@
 
 if __name__ == "__main__":
     
-    if len(sys.argv[:])!=4:
-        print "python mapSegToGeneMatrix.py genomicsSegmentIn refGene GeneLevelMatrixOut\n"
+    if len(sys.argv[:])!=5:
+        print "python mapSegToGeneMatrix.py genomicsSegmentIn refGene GeneLevelMatrixOut NORMAL_CNV\n"
         sys.exit()
     refgene = CGData.RefGene.RefGene()
     refgene.load( sys.argv[2] )
 
+    NORMAL_CNV=sys.argv[4]
+
     #* b for cnv 
     probeMapper = CGData.GeneMap.ProbeMapper('b')
 
@@ -74,7 +76,7 @@
         for sample in sample_list:
             list = matrix[samples[sample]][genes[gene]]
             if len(list)==0:
-                average =0
+                average = NORMAL_CNV
             elif len(list)==1:
                 average = list[0]
                 average =round(average,3)
--- a/seg2matrix/segToMatrixGalaxy.py	Thu Sep 17 15:38:53 2015 -0700
+++ b/seg2matrix/segToMatrixGalaxy.py	Thu Sep 17 22:03:04 2015 -0700
@@ -33,24 +33,25 @@
 
 
 if __name__ == "__main__":
-    if len(sys.argv[:])!= 5:
-        print "python segToMatrixGalaxy.py inputSegmentFile refGeneFile outputMatrix outputProbeMap\n"
+    if len(sys.argv[:])!= 6:
+        print "python segToMatrixGalaxy.py inputSegmentFile refGeneFile outputMatrix outputProbeMap NORMAL_CNV\n"
         sys.exit()
     seg = CGData.GenomicSegment.GenomicSegment()
     seg.load(sys.argv[1])
     
     refgene = CGData.RefGene.RefGene()
-    refgene.load(sys.argv[2])
-#    refgene.load(os.path.dirname(sys.argv[0])+"/"+sys.argv[2])
+    refgene.load(os.path.dirname(sys.argv[0])+"/"+os.path.basename(sys.argv[2]))
+
+    NORMAL_CNV = sys.argv[5]
 
     handle = open(sys.argv[3], "w")
     m = matrix_write(handle)    
-    CGData.SegToMatrix.seg_to_matrix(seg, m)
+    CGData.SegToMatrix.seg_to_matrix(seg, m, NORMAL_CNV)
     handle.close()
     
     handle = open(sys.argv[4], "w")
     probeMapper = CGData.GeneMap.ProbeMapper('b')
-    handle.write("%s\t%s\t%s\t%s\t%s\t%s\n" % ("#id", "gene","chrom","chromStart","chromEnd","strand"))
+    handle.write("%s\t%s\t%s\t%s\t%s\t%s\n" % ("id", "gene","chrom","chromStart","chromEnd","strand"))
     for probe in m.probes:
         hits = []
         for hit in probeMapper.find_overlap( probe, refgene ):
--- a/segToGeneMatrix.xml	Thu Sep 17 15:38:53 2015 -0700
+++ b/segToGeneMatrix.xml	Thu Sep 17 22:03:04 2015 -0700
@@ -3,7 +3,7 @@
     Convert segmented copy number data to gene-level matrix data
   </description>
   <command interpreter="python">
-    seg2matrix/mapSegToGeneMatrix.py $input $__tool_directory__/seg2matrix/$refGene.assembly $outputMatrix
+    seg2matrix/mapSegToGeneMatrix.py $input $__tool_directory__/seg2matrix/$refGene.assembly $outputMatrix $normal_CNV_value
   </command>
   <inputs>
     <param name="input" format="tabular" type="data" multiple="false" label="Input segmented copy number data" help="See input data format section for required input format" />
@@ -18,7 +18,8 @@
       <when value="hg18">
         <param name="assembly" type="hidden" value="refGene_hg18" />
       </when>
-    </conditional>  
+    </conditional>
+    <param name="normal_CNV_value" type="text" value="0" lable="what is the value to represent normal copy number" />
   </inputs>
   <outputs>
     <data name="outputMatrix" format="tabular" label="gene-level copy number matrix" />
--- a/segToMatrix.xml	Thu Sep 17 15:38:53 2015 -0700
+++ b/segToMatrix.xml	Thu Sep 17 22:03:04 2015 -0700
@@ -3,7 +3,7 @@
     Prepare segmented copy number data for Xena
   </description>
   <command interpreter="python">
-    seg2matrix/segToMatrixGalaxy.py $input $__tool_directory__/seg2matrix/$refGene.assembly $outputMatrix $outputProbeMap
+    seg2matrix/segToMatrixGalaxy.py $input $__tool_directory__/seg2matrix/$refGene.assembly $outputMatrix $outputProbeMap $normal_CNV_value
   </command>
   <inputs>
     <param name="input" format="tabular" type="data" multiple="false" label="Input segmented copy number data" />
@@ -18,7 +18,8 @@
       <when value="hg18">
         <param name="assembly" type="hidden" value="refGene_hg18" />
       </when>
-    </conditional>  
+    </conditional>
+    <param name="normal_CNV_value" type="text" value="0" lable="what is the value to represent normal copy number" />
   </inputs>
   <outputs>
     <data name="outputMatrix" format="tabular" label="xena copy number matrix" />