BAM format

The input/output sequencing files used by PacBio Data Processing are BAM files (see SAM/BAM format for a description).

The bam file

The bam file is is a binary file. It is a compressed version of a sam file —that is a human readable format— containing all the sequencing information. To manipulate the bam file format you can use packages like pysam compatible only with python 2 or pybam for python2 and pybam for python 3. However, for the PacBio bam file Pysam and Pybam are usefull to explore basic fiels in the aligned PacBio bam file.

PacBio bam file

Compared with the standard bam format, the PacBio.bam have extra columns as those containing the Interpulse Duration (IPD) value and Pulse Width (PW) that are important for the kinetic analysis.

The bam file of the main PacBio output (filename.subreads.bam) contains 26 columns which can be inspected using samtools. For instance, the following shell command tells us the number of columns in a bam file:

samtools view filename.subreads.bam|awk -F'\t' '{print NF; exit}'

In the next table, the most important columns in a PacBio bam are described:

Column number	Tag or content	Description
1	Molecule identifier	Molecule indentifier containing {movieName}/ {holeNumber}/ {qStart}_{qEnd}
10	…AGTAC…	Sequence
11	…~~C!~…	QUAL
12	RG	ReadGroup
13	dq	DeletionQV
14	dt	DeletionTag
15	ip	Ipd: B,C or B,S (raw frames or codec V1)
16	iq	InsertionQv
17	mq	MergeQv
18	np	NumPasses
19	pw	PulseWith: B,C or B,S (raw frames or codec V1)
20	qe	0_based end
21	qs	0_based start
22	rq	Float in [0,1] encoding expected accuracy
23	sn	4 floats for the average signal-to-noise ratio of A,C,G, and T (in that order) over the HQ region
24	sq	SubstitutionQV
25	zm	ZNW hole number
26	cx	Subread local context flags

If you want to check the differents length of the subreads using command line, you can type:

samtools view filename.subreads.bam|awk '{print length ($10)}'|sort -nur

For more information see the following link: BAM format specification for PacBio

Aligned PacBio bam file

The aligned bam file of the main PacBio output (moviename.subreads.bam) contains 28 columns. Again, the next shell one-liner counts the columns in the bam:

samtools view moviename.subreads.bam|awk -F'\t''{print NF; exit}'

It follows a brief description of the most important columns:

Column number	Tag or content	Description
1	Molecule identifier	Molecule indentifier containing {movieName}/ {holeNumber}/ {qStart}_{qEnd}
2	mapping flag	Value related to the alignment type (forward strand (0) and reverse strand (16) are the most important. More details in the link ‘’Map Format Specification’’ below)
4	position	Position where the sequence was mapped
5	mapping quality	Quality of the mapping
10	…AGTAC…	Sequence
11	…~~C!~…	QUAL
12	RG	ReadGroup
13	dq	DeletionQV
14	dt	DeletionTag
15	ip	Ipd: B,C or B,S (raw frames or codec V1)
16	iq	InsertionQv
17	mq	MergeQv
18	np	NumPasses
19	pw	PulseWith: B,C or B,S (raw frames or codec V1)
20	qe	0_based end
21	qs	0_based start
22	rq	Float in [0,1] encoding expected accuracy
23	sn	4 floats for the average signal-to-noise ratio of A,C,G, and T (in that order) over the HQ region
24	sq	SubstitutionQV
25	zm	ZNW hole number
26	cx	Subread local context flags
27	AS	Alignment score generated by aligner
28	NM	Number of differences (mismatches plus inserted and deleted bases) between the sequence and reference

For more information:

• Sequence Alignment/Map Format Specification

• Sequence Alignment/Map Optional Fields Specification

Fields

In this section we give details on some particular fields (columns) in a bam file.

Quailty of sequencing

In the SAM/BAM format specification it is declared that the 11-th column in the alignment section of BAM files is named QUAL, and it is described like follows:

(brief description) ASCII of Phred-scaled base QUALity+33

QUAL: ASCII of base QUALity plus 33 (same as the quality string in the Sanger FASTQ format). A base quality is the phred-scaled base error probability which equals -10 log10 Pr{base is wrong}. This field can be a ‘*’ when quality is not stored. If not a ‘*’, SEQ must not be a ‘*’ and the length of the quality string ought to equal the length of SEQ.

And the Wikipedia (FASTQ) explains:

The byte representing quality runs from 0x21 (lowest quality; ! in ASCII) to 0x7e (highest quality; ~ in ASCII). Here are the quality value characters in left-to-right increasing order of quality (ASCII):

!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~

For example, each base in a sequence like (in the 10-th column of a BAM file):

AATGCTAGCTAGCTCCTTGGATCGATCCGAT

will have an ASCII symbol (between ! and ~) associated with it that will be the contents of the 11-th column in the BAM file. For instance:

~~~~i~l~~~~_~~~~Z~~~~~~~~~~~~~~

Each symbol tells us the quality of sequencing the corresponding base.

Since the ASCII symbols ! and ~ correspond to 33 and 126 in decimal (or 0x21 and 0x7e in hexadecimal), and since each quality value is shifted by 33 it means that the range of allowed qualities, [0, 93], corresponds to a range of allowed probabilities for each base being wrong of, roughly [1, 0.00005] (beware the scale).