From d7596e7a08cbc56aaafd1c201e0e482f86590ba7 Mon Sep 17 00:00:00 2001
From: Ali Gholami Rudi <ali@rudi.ir>
Date: Tue, 16 Oct 2012 21:45:06 +0330
Subject: [PATCH] bl2exp: extract blast e-value from blastp output

---
 Makefile |  6 ++++--
 bl2exp.c | 71 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+), 2 deletions(-)
 create mode 100644 bl2exp.c

diff --git a/Makefile b/Makefile
index 499a2c2..80dba9a 100644
--- a/Makefile
+++ b/Makefile
@@ -2,7 +2,7 @@ CC = cc
 CFLAGS = -O2 -Wall
 LDFLAGS = -lpthread
 
-all: rangi conv genrq meval fastafilt
+all: rangi conv genrq meval fastafilt bl2exp
 %.o: %.c rangi.h
 	$(CC) -c $(CFLAGS) $<
 rangi: rangi.o mat.o proc.o
@@ -15,5 +15,7 @@ meval: meval.o
 	$(CC) -o $@ $^ $(LDFLAGS)
 fastafilt: fastafilt.o
 	$(CC) -o $@ $^ $(LDFLAGS)
+bl2exp: bl2exp.o
+	$(CC) -o $@ $^ $(LDFLAGS)
 clean:
-	rm -f *.o rangi conv genrq meval fastafilt
+	rm -f *.o rangi conv genrq meval fastafilt bl2exp
diff --git a/bl2exp.c b/bl2exp.c
new file mode 100644
index 0000000..7a9a22d
--- /dev/null
+++ b/bl2exp.c
@@ -0,0 +1,71 @@
+/*
+ * bl2exp, extract blast E-values from blastp output
+ *
+ * Copyright (C) 2012 Ali Gholami Rudi <ali at rudi dot ir>
+ *
+ * This program is released under the modified BSD license.
+ */
+#include <ctype.h>
+#include <stdio.h>
+#include <string.h>
+
+#define ETHRESH		0.001
+
+static char *readword(char *d, char *s)
+{
+	while (isspace(*s))
+		s++;
+	while (*s && !isspace(*s))
+		*d++ = *s++;
+	*d = '\0';
+	return s;
+}
+
+static int readentries(void)
+{
+	char line[1024];
+	char query[128];
+	char other[128];
+	char s_score[128];
+	int junk;
+	double score;
+	while (1) {
+		if (!fgets(line, sizeof(line), stdin))
+			return 1;
+		if (!strncmp(line, "Query= ", 6)) {
+			readword(query, line + 7);
+			break;
+		}
+	}
+	while (1) {
+		if (!fgets(line, sizeof(line), stdin))
+			return 1;
+		if (!strncmp(line, "Sequences producing", 16))
+			break;
+		if (!strncmp(line, " ***** No hits found", 16))
+			return 0;
+	}
+	fgets(line, sizeof(line), stdin);
+	while (1) {
+		if (!fgets(line, sizeof(line), stdin))
+			return 1;
+		if (sscanf(line, "%s %d %s", other, &junk, s_score) != 3)
+			break;
+		sprintf(line, "%s%s", s_score[0] == 'e' ? "1" : "", s_score);
+		sscanf(line, "%lf", &score);
+		if (score < ETHRESH)
+			printf("%s\t%s\t%lg\n", query, other, score);
+	}
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	if (argc > 1) {
+		printf("usage: %s <blast >e_values\n", argv[0]);
+		return 0;
+	}
+	while (!readentries())
+		;
+	return 0;
+}
-- 
2.11.4.GIT