Skip to main content

2.2 SAM/BAM Format

SAM (Sequence Alignment Map) and BAM (Binary Alignment Map) are the standard formats for storing aligned sequencing reads. BAM is the compressed binary version of SAM.

Overview

After aligning sequencing reads to a reference genome, the output is typically a SAM or BAM file containing:

  • Read sequences
  • Quality scores
  • Alignment positions
  • Mapping quality
  • Alignment flags
  • Optional tags with additional information

SAM Format Structure

SAM files are tab-delimited text files with two sections:

1. Header Section (Optional)

Lines starting with @:

@HD VN:1.6  SO:coordinate
@SQ SN:chr1 LN:248956422
@SQ SN:chr2 LN:242193529
@PG ID:bwa  PN:bwa  VN:0.7.17

Header tags: - @HD - Header line (version, sort order) - @SQ - Reference sequence dictionary - @PG - Program information - @RG - Read group information

2. Alignment Section

Each line represents one alignment:

READ1   99  chr1    10000   60  75M =   10200   275 ACGTACGTACGT... IIIIIIIIII...   NM:i:0  MD:Z:75 AS:i:75

SAM Fields

Col Field Description
1 QNAME Query (read) name
2 FLAG Bitwise flag (alignment properties)
3 RNAME Reference sequence name (chromosome)
4 POS 1-based leftmost mapping position
5 MAPQ Mapping quality (0-255)
6 CIGAR CIGAR string (alignment representation)
7 RNEXT Reference name of mate/next read
8 PNEXT Position of mate/next read
9 TLEN Template length (insert size)
10 SEQ Read sequence
11 QUAL Base quality scores (ASCII)
12+ TAGs Optional fields (e.g., NM, MD, AS)

Understanding SAM Flags

The FLAG field is a bitwise flag indicating alignment properties:

Common flags:

  • 0 - Read mapped, no special properties
  • 4 - Read unmapped
  • 16 - Read reverse strand
  • 83 - First in pair, mate reverse strand, proper pair
  • 99 - First in pair, mate reverse strand, proper pair (read forward)
  • 147 - Second in pair, read reverse strand, proper pair
  • 163 - Second in pair, mate reverse strand, proper pair (read forward)

Common Flag Combinations

# Check if read is unmapped
samtools view -f 4 file.bam

# Check if read is mapped
samtools view -F 4 file.bam

# Get properly paired reads
samtools view -f 2 file.bam

# Get reads on reverse strand
samtools view -f 16 file.bam

Decode flags online: Explain SAM Flags

Viewing SAM Files

You can view sam files using samtools tview -p which provides a text-based alignment viewer:

 samtools tview -p chr11:32224569-32251632 EB_CD71_E14_WT_Med1_ChIP.bam

You can also use the reference genome fasta file for better visualization of mismatches:

samtools tview -p chr11:32224569-32251632 EB_CD71_E14_WT_Med1_ChIP.bam -f /databank/igenomes/Mus_musculus/UCSC/mm39/Sequence/WholeGenomeFasta/genome.fasta

If you want to exit tview, just press q.

If you are looking for variants the tview is a great way to visualize them quickly.

Viewing BAM Files in IGV

IGV (Integrative Genomics Viewer) is a popular tool for visualizing BAM files: See the IGV User Guide for instructions on loading BAM files and navigating the genome.

CIGAR String

CIGAR describes how the read aligns to the reference:

Operations:

  • M - Match or mismatch
  • I - Insertion (gap in reference)
  • D - Deletion (gap in read)
  • S - Soft clipping (clipped but sequence present)
  • H - Hard clipping (clipped and sequence removed)
  • N - Skipped region (for RNA-seq splicing)
  • = - Sequence match
  • X - Sequence mismatch

Example CIGAR strings:

  • 75M - 75 bases match/mismatch
  • 50M1I24M - 50 match, 1 insertion, 24 match
  • 40M5D35M - 40 match, 5 deletion, 35 match
  • 10S65M - 10 soft-clipped, 65 match
  • 25M100N50M - 25 match, 100 skipped (intron), 50 match (RNA-seq)

Optional Tags

Common optional tags:

  • NM:i:2 - Edit distance to reference
  • MD:Z:75 - String for mismatching positions
  • AS:i:75 - Alignment score
  • XS:i:70 - Suboptimal alignment score
  • RG:Z:sample1 - Read group identifier

BAM Format

BAM is the compressed binary version of SAM:

Advantages:

  • ~4x smaller file size
  • Faster to process
  • Supports indexing for rapid region queries

Disadvantages:

  • Not human-readable
  • Requires tools to view

Working with SAM/BAM Files

View BAM as SAM

samtools view file.bam | head

View BAM with header

samtools view -h file.bam | head

Convert SAM to BAM

samtools view -b input.sam -o output.bam

Convert BAM to SAM

samtools view -h input.bam -o output.sam

Sort BAM by coordinate

samtools sort input.bam -o sorted.bam

Index BAM file

samtools index sorted.bam
# Creates sorted.bam.bai

Extract specific region

# Requires indexed BAM
samtools view sorted.bam chr1:10000-20000

Filter alignments

# Only mapped reads
samtools view -F 4 -b input.bam -o mapped.bam

# Only properly paired reads
samtools view -f 2 -b input.bam -o proper_pairs.bam

# Minimum mapping quality 30
samtools view -q 30 -b input.bam -o high_quality.bam

Count reads

# Total reads
samtools view -c file.bam

# Mapped reads
samtools view -c -F 4 file.bam

# Unmapped reads
samtools view -c -f 4 file.bam

BAM File Statistics

Basic stats with samtools

samtools flagstat file.bam

Output includes:

  • Total reads
  • Mapped reads
  • Properly paired reads
  • Singletons
  • Duplicates

Detailed stats

samtools stats file.bam > stats.txt

Best Practices

  1. Always sort and filter BAM files before indexing or downstream analysis
  2. Index sorted BAM files (.bam.bai) for efficient access
  3. Delete the SAM and only keep BAM (and .bam.bai) format for storage (smaller, faster)

File Size Comparison

For a typical experiment:

  • FASTQ: 10 GB
  • SAM: 15 GB
  • BAM (unsorted): 4 GB
  • BAM (sorted): 3 GB

Additional Resources