Implemented functions to build reverse complement sequences

src/utils.c

@@ -28,6 +28,140 @@
 #define DEBUG_LEVEL 0	// TODO has to be defined somewhere else (cython compil flag?)
 
 
+/**************************************************************************
+ *
+ * D E C L A R A T I O N   O F   T H E   P R I V A T E   F U N C T I O N S
+ *
+ **************************************************************************/
+
+
+/**
+ * Internal function returning the complement of a nucleotide base.
+ *
+ * @warning The base must be in lower case.
+ *
+ * @param nucAc The nucleotide base.
+ *
+ * @returns The complement of the nucleotide base.
+ * @retval The nucleotide base itself if no complement was found.
+ *
+ * @since December 2016
+ * @author Celine Mercier (celine.mercier@metabarcoding.org)
+ * @note Copied from ecoPCR source code
+ */
+static char nuc_base_complement(char nucAc);
+
+
+/**
+ * Internal function returning the complement of a nucleotide sequence.
+ *
+ * @warning The sequence must be in lower case.
+ * @warning The sequence will be replaced by its complement without being copied.
+ *
+ * @param nucAcSeq The nucleotide sequence.
+ *
+ * @returns The complemented sequence.
+ *
+ * @since December 2016
+ * @author Celine Mercier (celine.mercier@metabarcoding.org)
+ * @note Copied from ecoPCR source code
+ */
+static char* nuc_seq_complement(char* nucAcSeq);
+
+
+/**
+ * Internal function returning the reverse of a nucleotide sequence.
+ *
+ * @warning The sequence must be in lower case.
+ * @warning The sequence will be replaced by its reverse without being copied.
+ *
+ * @param str The nucleotide sequence.
+ * @param isPattern Whether the sequence is a pattern. TODO
+ *
+ * @returns The reversed sequence.
+ *
+ * @since December 2016
+ * @author Celine Mercier (celine.mercier@metabarcoding.org)
+ * @note Copied from ecoPCR source code
+ */
+static char* reverse_sequence(char* str, char isPattern);
+
+
+/************************************************************************
+ *
+ * D E F I N I T I O N   O F   T H E   P R I V A T E   F U N C T I O N S
+ *
+ ************************************************************************/
+
+static char nuc_base_complement(char nucAc)
+{
+    char* c;
+
+    if ((c = strchr(DNA_ALPHA, nucAc)))
+        return CDNA_ALPHA[(c - DNA_ALPHA)];
+    else
+        return nucAc;
+}
+
+
+static char* nuc_seq_complement(char* nucAcSeq)
+{
+    char *s;
+
+    for (s = nucAcSeq ; *s ; s++)
+        *s = nuc_base_complement(*s);
+
+    return nucAcSeq;
+}
+
+
+static char* reverse_sequence(char* str, char isPattern)
+{
+        char *sb, *se, c;
+
+        if (! str)
+            return str;
+
+        sb = str;
+        se = str + strlen(str) - 1;
+
+        while(sb <= se) {
+           c    = *sb;
+          *sb++ = *se;
+          *se-- = c;
+        }
+
+		sb = str;
+		se = str + strlen(str) - 1;
+
+		if (isPattern)
+			for (;sb < se; sb++)
+			{
+				if (*sb=='#')
+				{
+					if (((se - sb) > 2) && (*(sb+2)=='!'))
+					{
+						*sb='!';
+						sb+=2;
+						*sb='#';
+					}
+					else
+					{
+						*sb=*(sb+1);
+						sb++;
+						*sb='#';
+					}
+				}
+				else if (*sb=='!')
+					{
+						*sb=*(sb-1);
+						*(sb-1)='!';
+					}
+			}
+
+        return str;
+}
+
 
 /**********************************************************************
  *
@@ -404,3 +538,15 @@ loop:   SWAPINIT(a, es);
 /*              qsort(pn - r, r / es, es, cmp);*/
 }
 
+
+char* reverse_complement_pattern(char* nucAcSeq)
+{
+    return reverse_sequence(nuc_seq_complement(nucAcSeq), 1);
+}
+
+
+char* reverse_complement_sequence(char* nucAcSeq)
+{
+    return reverse_sequence(nuc_seq_complement(nucAcSeq), 0);
+}
+