src/hg/instinct/extractData/extractData.c 1.5
1.5 2010/01/14 23:15:04 jsanborn
updated
Index: src/hg/instinct/extractData/extractData.c
===================================================================
RCS file: /projects/compbio/cvsroot/kent/src/hg/instinct/extractData/extractData.c,v
retrieving revision 1.4
retrieving revision 1.5
diff -b -B -U 4 -r1.4 -r1.5
--- src/hg/instinct/extractData/extractData.c 14 Jan 2010 07:12:28 -0000 1.4
+++ src/hg/instinct/extractData/extractData.c 14 Jan 2010 23:15:04 -0000 1.5
@@ -26,17 +26,22 @@
" extractData [OPTIONS] db table sample_id genes(s)\n"
"options:\n"
" -median Output median value if multiple probes\n"
" -tcga Handle TCGA ids, keeps only first 16 chars in sample_id\n"
+ " -samples=STR Sample comma-separated list, otherwise all sample data are returned.\n"
);
}
+#define TCGA_PATIENT_PREFIX 12
+#define TCGA_SAMPLE_PREFIX 16
+
boolean median = FALSE;
boolean isTCGA = FALSE;
static struct optionSpec options[] = {
{"median", OPTION_BOOLEAN},
{"tcga", OPTION_BOOLEAN},
+ {"samples", OPTION_STRING},
{NULL, 0},
};
struct maGrouping *getMaGrouping(struct sqlConnection *hgConn, char *tableName)
@@ -130,9 +135,18 @@
return gaHash;
}
-struct slDouble *getProbeData(struct sqlConnection *hgConn, int expId,
+struct sampleVals {
+ struct sampleVals *next;
+
+ int expId;
+ char *name;
+ struct slDouble *vals;
+ struct hash *valHash;
+};
+
+void setProbeData(struct sqlConnection *hgConn, struct sampleVals *svList,
char *dataTable, struct slName *probes)
{
struct slName *sl;
struct dyString *dy = newDyString(100);
@@ -151,34 +165,139 @@
/* Get bed15 data from hg18 database */
struct sqlResult *sr = sqlGetResult(hgConn, query);
char **row = NULL;
-struct slDouble *sdList = NULL;
+struct sampleVals *sv;
while ((row = sqlNextRow(sr)) != NULL)
{
struct bed *nb = bedLoadN(row+1, 15);
- if (expId >= nb->expCount)
+ for (sv = svList; sv; sv = sv->next)
+ {
+ if (sv->expId >= nb->expCount)
continue;
-
- struct slDouble *sd = slDoubleNew(nb->expScores[expId]);
- slAddHead(&sdList, sd);
+ struct slDouble *sd = slDoubleNew(nb->expScores[sv->expId]);
+ slAddHead(&sv->vals, sd);
+ hashAdd(sv->valHash, nb->name, sd);
+ }
bedFree(&nb);
}
sqlFreeResult(&sr);
+}
+
+
+struct sampleVals *prepareSampleVals(char *samples, struct maGrouping *allA)
+{
+boolean allSamples = FALSE;
+struct hash *sampleHash = hashNew(0);
+
+if (!samples)
+ allSamples = TRUE;
+else
+ {
+ struct slName *sl, *slList = slNameListFromComma(samples);
+ for (sl = slList; sl; sl = sl->next)
+ {
+ char *sample = sl->name;
+ if (isTCGA)
+ sample = cloneStringZ(sl->name, TCGA_SAMPLE_PREFIX);
+ hashAddInt(sampleHash, sample, 1);
+ }
+ }
+
+struct hashEl *el;
+struct sampleVals *sv, *svList = NULL;
+int i;
+for (i = 0; i < allA->size; i++)
+ {
+ int expId = allA->expIds[i];
+ char *maName = allA->names[i];
+ if (isTCGA)
+ maName = cloneStringZ(allA->names[i], TCGA_SAMPLE_PREFIX);
+
+ if (allSamples || (el = hashLookup(sampleHash, maName)) != NULL)
+ {
+ sv = AllocA(struct sampleVals);
+ sv->name = cloneString(allA->names[i]);
+ sv->expId = expId;
+ sv->vals = NULL;
+ sv->valHash = hashNew(0);
+
+ slAddHead(&svList, sv);
+ }
+
+ }
+slReverse(&svList);
+
+return svList;
+}
+
+void printProbeData(struct geneAlias *ga, struct sampleVals *svList)
+{
+struct sampleVals *sv;
+struct dyString *dy = newDyString(100);
+
+if (median)
+ {
+ dyStringPrintf(dy, "%s\tmedian\t", ga->gene);
+ for (sv = svList; sv; sv = sv->next)
+ {
+ double val = slDoubleMedian(sv->vals);
+ dyStringPrintf(dy, "%f", val);
+
+ if (sv->next)
+ dyStringPrintf(dy, "\t");
+ }
+ dyStringPrintf(dy, "\n");
+ }
+else
+ {
+ struct hashEl *el;
+ struct slName *sl;
+ for (sl = ga->probes; sl; sl = sl->next)
+ {
+ dyStringPrintf(dy, "%s\t%s\t", ga->gene, sl->name);
+ for (sv = svList; sv; sv = sv->next)
+ {
+ if ((el = hashLookup(sv->valHash, sl->name)) != NULL)
+ {
+ struct slDouble *sd = el->val;
+ dyStringPrintf(dy, "%f", sd->val);
+ }
+ else
+ dyStringPrintf(dy, "NA");
+
+ if (sv->next)
+ dyStringPrintf(dy, "\t");
+ }
+ dyStringPrintf(dy, "\n");
+ }
+ }
-return sdList;
+char *valStr = dyStringCannibalize(&dy);
+printf("%s", valStr);
}
-void extractData(char *tableName, char *sampleName, char *geneList)
+void clearProbeData(struct sampleVals *svList)
+{
+struct sampleVals *sv;
+for (sv = svList; sv; sv = sv->next)
+ {
+ slFreeList(&sv->vals);
+ hashFree(&sv->valHash);
+
+ sv->vals = NULL;
+ sv->valHash = hashNew(0);
+ }
+}
+
+
+void extractData(char *tableName, char *geneList, char *samples)
{
if (!geneList)
return;
-char *sample = sampleName;
-if (isTCGA)
- sample = cloneStringZ(sampleName, 16);
struct slName *genes = slNameListFromComma(geneList);
struct hashEl *el;
@@ -188,25 +307,9 @@
struct maGrouping *allA = getMaGrouping(hgConn, tableName);
if (!allA)
errAbort("Could not find maGrouping for %s!", tableName);
-int i = 0, expId = -1;
-for (; i < allA->size; i++)
- {
- char *maName = allA->names[i];
- if (isTCGA)
- maName = cloneStringZ(allA->names[i], 16);
-
- if (!sameString(maName, sample))
- continue;
- expId = allA->expIds[i];
- break;
- }
-if (expId < 0)
- errAbort("Couldn't find sample %s in %s.\n", sample, tableName);
-
struct hash *settings = getSettings(tableName);
-
el = hashLookup(settings, "aliasTable");
if (!el)
errAbort("No aliasTable.\n");
char *aliasTable = cloneString(el->val);
@@ -212,37 +315,35 @@
char *aliasTable = cloneString(el->val);
struct hash *gaHash = getAliases(hgConn, aliasTable, genes);
+struct sampleVals *sv, *svList = prepareSampleVals(samples, allA);
+if (!svList)
+ errAbort("No samples allocated.\n");
+
+printf("gene\tprobe");
+for (sv = svList; sv; sv = sv->next)
+ printf("\t%s", sv->name);
+printf("\n");
+
struct hashCookie cookie = hashFirst(gaHash);
while ((el = hashNext(&cookie)) != NULL)
{
- char *name = el->name;
struct geneAlias *ga = el->val;
- struct slDouble *sd, *sdList = getProbeData(hgConn, expId, tableName, ga->probes);
+ setProbeData(hgConn, svList, tableName, ga->probes);
+
+ printProbeData(ga, svList);
+
+ clearProbeData(svList);
+ }
+#if 0
if (!sdList)
continue;
- printf("%s\t", name);
- if (median)
- {
- double val = slDoubleMedian(sdList);
- printf("%f\n", val);
- }
- else
- {
- struct dyString *dy = newDyString(100);
- for (sd = sdList; sd; sd = sd->next)
- {
- dyStringPrintf(dy, "%f", sd->val);
- if (sd->next)
- dyStringPrintf(dy, ",");
- }
- char *val = dyStringCannibalize(&dy);
- printf("%s\n", val);
- }
+
}
+#endif
hFreeConn(&hgConn);
}
@@ -250,15 +351,16 @@
int main(int argc, char *argv[])
/* Process command line. */
{
optionInit(&argc, argv, options);
-if (argc != 4)
+if (argc != 3)
usage();
if (optionExists("median"))
median = TRUE;
if (optionExists("tcga"))
isTCGA = TRUE;
-extractData(argv[1], argv[2], argv[3]);
+char *samples = optionVal("samples", NULL);
+extractData(argv[1], argv[2], samples);
return 0;
}