Convert U to T instead of U to N when sam_parsing.

As this changes seq_nt16_table it changes it for all other uses of
this lookup table too, which will be widespread.  However this feels
like a reasonable thing to do given it only has an impact on data
which is currently out of bounds of what is expected.

Fixes samtools/samtools#2131

Author: jkbonfield

hts.c

@@ -232,16 +232,19 @@ const char *hts_feature_string(void) {
 }
 
 
+// Converts ASCII to BAM nibble encoding.
+// Note 0123 is treated as ACGT (ABI colourspace encoding) and
+// U is treated as T.
 HTSLIB_EXPORT
 const unsigned char seq_nt16_table[256] = {
     15,15,15,15, 15,15,15,15, 15,15,15,15, 15,15,15,15,
     15,15,15,15, 15,15,15,15, 15,15,15,15, 15,15,15,15,
     15,15,15,15, 15,15,15,15, 15,15,15,15, 15,15,15,15,
      1, 2, 4, 8, 15,15,15,15, 15,15,15,15, 15, 0 /*=*/,15,15,
     15, 1,14, 2, 13,15,15, 4, 11,15,15,12, 15, 3,15,15,
-    15,15, 5, 6,  8,15, 7, 9, 15,10,15,15, 15,15,15,15,
+    15,15, 5, 6,  8, 8, 7, 9, 15,10,15,15, 15,15,15,15,
     15, 1,14, 2, 13,15,15, 4, 11,15,15,12, 15, 3,15,15,
-    15,15, 5, 6,  8,15, 7, 9, 15,10,15,15, 15,15,15,15,
+    15,15, 5, 6,  8, 8, 7, 9, 15,10,15,15, 15,15,15,15,
 
     15,15,15,15, 15,15,15,15, 15,15,15,15, 15,15,15,15,
     15,15,15,15, 15,15,15,15, 15,15,15,15, 15,15,15,15,

htslib/hts.h

@@ -455,6 +455,7 @@ int hts_parse_opt_list(htsFormat *opt, const char *str);
 The input character may be either an IUPAC ambiguity code, '=' for 0, or
 '0'/'1'/'2'/'3' for a result of 1/2/4/8.  The result is encoded as 1/2/4/8
 for A/C/G/T or combinations of these bits for ambiguous bases.
+Additionally RNA U is treated as a T (8).
 */
 HTSLIB_EXPORT
 extern const unsigned char seq_nt16_table[256];

test/compare_sam.pl

@@ -163,6 +163,9 @@
     $ln1[9] = uc($ln1[9]);
     $ln2[9] = uc($ln2[9]);
 
+    # RNA U to T is an expected change
+    $ln1[9] =~ s/U/T/g;
+
     # Cram will populate a sequence string that starts as "*"
     $ln2[9] = "*" if ($ln1[9] eq "*");
 

test/xx#u.sam

@@ -0,0 +1,3 @@
+@SQ	SN:xx	LN:20
+a1	99	xx	1	1	16M	=	11	20	=ACMGRSVTWYHKDBN	****************
+b1	99	xx	1	1	16M	=	11	20	=ACMGRSVUWYHKDBN	****************