January 15, 2026 · 3 min read · By Smallest.zip Team

FASTQ vs BAM: Which Is Cheaper to Store?

Comparing FASTQ and BAM file sizes, compression ratios, and cloud storage costs. How to minimize genomics storage spend across both formats.

Compression FASTQ BAM Genomics Storage S3 Cloud

FASTQ vs BAM: The Storage Trade-Off

Every sequencing pipeline produces both FASTQ (raw reads) and BAM (aligned reads). Most organizations keep both — FASTQ for reprocessing and BAM for downstream analysis. But which one costs more to store, and how much can you save by compressing each?

File Size Comparison

For a typical 30x whole-genome sample:

Format	Typical Size	What It Contains
FASTQ (paired)	200 GB	Raw reads + quality scores
BAM (aligned)	120 GB	Aligned reads + CIGAR + metadata
CRAM 3.1	~50 GB	Reference-based BAM compression

FASTQ is larger because it stores redundant quality score data. BAM is smaller but contains alignment information. Many labs store both, totaling 320 GB per sample.

Compression Ratios

Here's how our compressor handles each format:

Format	Raw	gzip	Our Compression	Ratio
FASTQ	200 GB	50 GB	9 GB	4.5%
BAM	120 GB	—	17 GB	14.3%
Total	320 GB	170 GB	26 GB	8.1%

From 320 GB down to 26 GB per sample — a 92% reduction.

Cost at Scale

For 10,000 whole-genome samples on S3:

Storage Strategy	Total Size	Annual S3 Cost
FASTQ + BAM (uncompressed)	3.2 PB	$883,200
FASTQ (gzip) + BAM	1.7 PB	$469,200
FASTQ (4BIN) + BAM (compressed)	260 TB	$71,760

That's $811,440/year saved by compressing both formats vs storing uncompressed.

Should You Keep Both?

Keep FASTQ if:

You may need to re-align with a newer reference genome
Your pipeline requires raw reads for custom processing
Regulatory requirements mandate raw data retention

Keep only BAM if:

Your alignment is final and won't be re-run
You need fast access to aligned reads for variant calling
Storage is your primary bottleneck

Our recommendation: Keep both, compressed. At 26 GB per sample (compressed FASTQ + BAM), you can store 10,000 genomes for $71,760/year on S3 — less than what most labs pay for uncompressed BAM alone.

Getting Started

Compress both FASTQ and BAM files through our API. Both formats are fully lossless — the decompressed output is bit-identical to the original.