Confused about the strand FLAG of Bismark paired-end seq results

Jerry_Zhao · 09-13-2012, 02:11 PM

Can anyone help me about the strand FLAG in the Bismark results for paired-end data?

Previously, I was working with single-end seq analysis using Bismark. It works perfectly for me.
Nevertheless, recently I am working on some paired-end data. I am confused about the strand FLAG in the Bismark results.

My paired-end data are directional. Therefore,
All "OT" have FLAG number "67"
All "CTOT" have FLAG number "131"
All "OB" have FLAG number "115"
All "CTOB" have FLAG number "179"

Correct me if I am wrong:
when FLAG is "67" or "131", it means the reads on the "+" strand?
The strand used as the genomic sequences for mouse.

When FLAG is "115" or "179", it means the reads on the "-" strand?

Am I right?

fkrueger · 09-14-2012, 03:09 AM

Quote:

Originally Posted by Jerry_Zhao

Can anyone help me about the strand FLAG in the Bismark results for paired-end data?

Previously, I was working with single-end seq analysis using Bismark. It works perfectly for me.
Nevertheless, recently I am working on some paired-end data. I am confused about the strand FLAG in the Bismark results.

My paired-end data are directional. Therefore,
All "OT" have FLAG number "67"
All "CTOT" have FLAG number "131"
All "OB" have FLAG number "115"
All "CTOB" have FLAG number "179"

Correct me if I am wrong:
when FLAG is "67" or "131", it means the reads on the "+" strand?
The strand used as the genomic sequences for mouse.

When FLAG is "115" or "179", it means the reads on the "-" strand?

Am I right?

This is indeed a very confusing issues, I will link the commented code from the paired-end SAM section in the hope it answers your questions:

### As the FLAG value do not consider that there might be 4 different bisulfite strands of DNA, we are trying to make FLAG tags which take the strand identity into account

# strands OT and CTOT will be treated as aligning to the top strand (both sequences are scored as aligning to the top strand)
# strands OB and CTOB will be treated as aligning to the bottom strand (both sequences are scored as reverse complemented sequences)

Code:

### This is a description of the bitwise FLAG field which needs to be set for the SAM file taken from: "The SAM Format Specification (v1.4-r985), September 7, 2011"
  ## FLAG: bitwise FLAG. Each bit is explained in the following table:
  ## Bit    Description                                                Comment                                Value
  ## 0x1    template having multiple segments in sequencing            0: single-end 1: paired end            value: 2^^0 (  1)
  ## 0x2    each segment properly aligned according to the aligner     true only for paired-end alignments    value: 2^^1 (  2)
  ## 0x4    segment unmapped                                           ---                                           ---
  ## 0x8    next segment in the template unmapped                      ---                                           ---
  ## 0x10   SEQ being reverse complemented                             - strand alignment                     value: 2^^4 ( 16)
  ## 0x20   SEQ of the next segment in the template being reversed     + strand alignment                     value: 2^^5 ( 32)
  ## 0x40   the first segment in the template                          read 1                                 value: 2^^6 ( 64)
  ## 0x80   the last segment in the template                           read 2                                 value: 2^^7 (128)
  ## 0x100  secondary alignment                                        ---                                           ---
  ## 0x200  not passing quality controls                               ---                                           ---
  ## 0x400  PCR or optical duplicate                                   ---                                           ---


  if ($index == 0){       # OT
    $flag_1 = 67;                                                      # Read 1 is on the + strand  (1+2+64) (Read 2 is technically reverse-complemented, but we do not score it)
    $flag_2 = 131;                                                     # Read 2 is on - strand but informative for the OT        (1+2+128)
  }
  elsif ($index == 1){    # CTOB
    $flag_1 = 115;                                                     # Read 1 is on the + strand, we score for OB  (1+2+16+32+64)
    $flag_2 = 179;                                                     # Read 2 is on the - strand  (1+2+16+32+128)
  }
  elsif ($index == 2){    # CTOT
    $flag_1 = 67;                                                      # Read 1 is on the - strand (CTOT) strand, but we score it for OT (1+2+64)
    $flag_2 = 131;                                                     # Read 2 is on the + strand, score it for OT (1+2+128)
  }
  elsif ($index == 3){    # OB
    $flag_1 = 115;                                                     # Read 1 is on the - strand, we score for OB  (1+2+16+32+64)
    $flag_2 = 179;                                                     # Read 2 is on the + strand  (1+2+16+32+128)
  }

swbarnes2 · 09-14-2012, 10:45 AM

67 = 64+2+1

That means the read came from the first fastq file, the mate is not reverse direction, the read is not in the reverse direction, the mate mapped, the read mapped, the reads are properly paired, and the reads are paired.

So yes, 67 means the read and its mate mapped in the forward direction, and the read came from the first fastq file. 131 means that the read and its mate mapped in the forward direction, and the read came from the second fastq file.

115 = 64+32+16+2+1

So that's first fastq file, mate is reversed, read is reversed, properly paired, paired.

same with 179, except those came from the second fastq file.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
.wig files for strand-specific paired-end RNA-Seq	ulitskyi	RNA Sequencing	4	06-12-2012 11:27 PM
how to determine strand from tophat output for paired-end RNA-seq data	jay2008	Bioinformatics	1	05-30-2012 04:46 AM
Bismark paired-end read orientation	xinyu	Epigenetics	4	03-16-2012 05:57 AM
Bismark paired-end positions	mixter	Epigenetics	1	01-19-2012 10:45 AM
bfast paired end flag	guavajuice	Bioinformatics	0	11-02-2010 11:37 AM

09-13-2012, 02:11 PM	#1
Jerry_Zhao Member Location: Philadelphia Join Date: Feb 2012 Posts: 20	Confused about the strand FLAG of Bismark paired-end seq results Can anyone help me about the strand FLAG in the Bismark results for paired-end data? Previously, I was working with single-end seq analysis using Bismark. It works perfectly for me. Nevertheless, recently I am working on some paired-end data. I am confused about the strand FLAG in the Bismark results. My paired-end data are directional. Therefore, All "OT" have FLAG number "67" All "CTOT" have FLAG number "131" All "OB" have FLAG number "115" All "CTOB" have FLAG number "179" Correct me if I am wrong: when FLAG is "67" or "131", it means the reads on the "+" strand? The strand used as the genomic sequences for mouse. When FLAG is "115" or "179", it means the reads on the "-" strand? Am I right?

09-14-2012, 10:45 AM	#3
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	67 = 64+2+1 That means the read came from the first fastq file, the mate is not reverse direction, the read is not in the reverse direction, the mate mapped, the read mapped, the reads are properly paired, and the reads are paired. So yes, 67 means the read and its mate mapped in the forward direction, and the read came from the first fastq file. 131 means that the read and its mate mapped in the forward direction, and the read came from the second fastq file. 115 = 64+32+16+2+1 So that's first fastq file, mate is reversed, read is reversed, properly paired, paired. same with 179, except those came from the second fastq file.