Read direction lost with BWA in SAM output?

[jorvis]

I've tried two different header styles in my input FASTQ headers when running BWA:

@SN7001163:162:C4A1UACXX:1:1101:1062:2076/1

and

@SN7001163:162:C4A1UACXX:1:1101:1062:2076 1:N:0:GTCCGCA

My goal is to be able to tell which mate I'm looking at in the FASTQ file, but it seems to get stripped in the SAM output, where from "bwa sampe" I get lines like this:

SN7001163:162:C4A1UACXX:1:1101:1062:2076	77	*	0	0	*	*	0	0	GTTTGCTTGGCTGTGAGCTTGTCCGACACGGGCCACCAGGAGAGTGAGATACACCGAGACGAGCATCCTGTCTTTCTCTCGGACGGTTCCACAACAAATAA	@?@DDD?;<;F>?<2A<E<FFC9:FE8):8@?FFF@FF=;=;D;).).7>77==EB'93;;3=@@:@(:3,+(4::@B>@5?-<@B<?<34>ABB1<8:43
SN7001163:162:C4A1UACXX:1:1101:1062:2076	141	*	0	0	*	*	0	0	GCCATGTTGAGTGAGAATTTATTATTTGTTGTGGAACC	;<;;(42@9)@)84):46=69416)2@:@:<=1(66@?

How can I tell which of these alignment lines refers to which input mate?

[jorvis]

I realize that those two reads didn't actually align, so the SAM lines were pretty minimal. Here are a pair which did:

SN7001163:162:C4A1UACXX:1:1101:1174:2116        81      Locus_14841_Transcript_1__1_Confidence_0.750_Length_603 292     37      101M    =       294     -99     CTCGTCATTTCAATGCCCCCTCTCATATCAGAAGGAAAATCATGAGTGCTCCTTTGTCAAAAGAGCTGAGAGCAAAGTACAATGTGAGAAGTATGCCCATT   >BBDDDDDDDDDDDBDFFHHHHIIHJJJJJJJJJJJJJJIIJJJJJJJJJJJJJJIIIJJJJIJJJJJJJIJJJJJJJHJJJIJJJJJHHHHHFFFFFCCC   XT:A:U  NM:i:0  SM:i:37 AM:i:37 X0:i:1  X1:i:0  XM:i:0  XO:i:0  XG:i:0  MD:Z:101
SN7001163:162:C4A1UACXX:1:1101:1174:2116        161     Locus_14841_Transcript_1__1_Confidence_0.750_Length_603 294     37      101M    =       292     99      CGTCATTTCAATGCCCCCTCTCATATCAGAAGGAAAATCATGAGTGCTCCTTTGTCAAAAGAGCTGAGAGCAAAGTACAATGTGAGAAGTATGCCCATTAG   BCBFFFFFHHHH?HIJJJJJJJJJJJJJJJJJJJJJJJJJJJJJHIJJJJGHIIHHIJJJJJJJJJIJJJJJHHHHHHFFFFFFFEEEEEEEEDDDDDDDC   XT:A:U  NM:i:1  SM:i:37 AM:i:37 X0:i:1  X1:i:0  XM:i:1  XO:i:0  XG:i:0  MD:Z:99C1

[WhatsOEver]

[original post deleted because I misunderstood the question]

gingers answer below is correct. You can simply confirm this by swapping R1 and R2 reads.

[gringer]

There are flags in the SAM file for the first (and last) read attached to a template sequence. If a bitwise and of the flag field with 0x40 returns non-zero, then it is the first read of a template sequence. In the case of the two examples you have, here is the full flag breakdown:

81 = 0101 0001
             Paired
        Reverse-complemented
      First read in the template

161 = 1010 0001
              Paired
        Other read is reverse-complemented
      Last read in the template

See https://samtools.github.io/hts-specs/SAMv1.pdf

These flags can be filtered using samtools view:

samtools view -b -f 0x40 in.bam > out_FirstRead.bam
samtools view -b -F 0x40 in.bam > out_notFirst.bam

The distinction between "last" and "second" is not important for most purposes, but there are some situations where more than two reads can be associated with the same template sequence.

Whether or not a read is first or last is particularly important for strand-specific sequencing, because it allows you to distinguish between templates that are oriented in the same direction as the primary transcript, and those that are not (e.g. siRNA).