Samtools Tview Legend

Question

[Pascal]

Hi.

I would like to understand the output screen produced by a "samtools tview". I may not googling with good keywords, but I just can't find any document explaining the meaining of ".", "," underlined characters, etc.

Thanks in advance.

Answer 1

[Pierre Lindenbaum]

See the C in samtools /bam_tview.c

c = bam1_strand(p->b)? ',' : '.'

• . is for a matching base reverse strand
• , is for a matching base forward strand

if (((p->b->core.flag&BAM*FPAIRED) && !(p->b->core.flag&BAM*FPROPER*PAIR)) || (p->b->core.flag & BAM*FSECONDARY)) attr |= A_UNDERLINE;

• an underline tells if the read was correctly paired or/not.
• etc...

[Pascal]

Oh it's clear now :->

[Pascal]

Thank you Pierre!

Answer 2

[Puriney]

I disagree with @Pierre.

, : negative strand,

. : positive strand.

You could try to separate the reads by its strand using the bitwise flag.

samtools -f 0x10 -b aln.bam> aln.neg.bam would give you all the reads mapped to negative strand,

samtools -F 0x10 -b aln.bam>aln.pos.bam would give the positive reads.

Then you could use samtools tview to see them. Due to alternative splicing, the RNA-seq reads would normally have 123N, for example, in their cigar string, which means skipped reads, a.k.a splicing junction spanning reads.

You would see the following for positive strand:

CATCACTGGTTTAAAGACAAACTTGCATTGTGAGATTCCAAAATAACAACAACAAAAAACAATTTGCATTGAGAACATTTTGAAG

.........A.......
.....>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
.....>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

You would see for negative strand.

TTTCATTTGCAAGTAATCGATTTAGGTTTTTGATTTTAGGGTTTTTTTTTGTTTTGAACAGTCCAGTCAAAGTACAAATCGAGAG
...KK....KKK..KK.K.K...K........K....K..................KKKK.........K...K...........
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<,,,,,,,,t,,,c,,,,,,,,,,

[Puriney]

The strand info comes from the aligner. Aligner get input from genome.fasta file to build index, meanwhile the genome.fasta file is recorded in 5'->3' direction. Thus if the read matches the raw genome.fasta sequence, the read is considered mapped on the positive strand, and vice versa. I hope so.