Compress
High-density lossless compression for highly structured scientific datasets.
ZCaps for genomics
Up to 99.26% less genomic data.
Lossless. Selectively accessible.
Measured on an 11.23 GiB VCF chr21 dataset using ZCaps Level 12. Original data reconstructed exactly.
Historical genomics campaign · results specific to the tested dataset, build and hardware.
VCF chr21ZCAPS · LEVEL 12
11.23 GiBSource dataset
ZCaps lossless compression
~85 MiBCompressed size, calculated
99.26%REDUCTION
134.6× compression ratio
Targeted 4 MiB extraction0.16–0.17 s
Lossless compression for genomic workflows
Access
Retrieve targeted byte ranges by decoding only the relevant compressed chunks.
Scale
Independent chunks and files can be processed concurrently across server resources.
ZCaps addresses bytes and ranges. Chromosome or genomic-coordinate lookup requires an integration layer that maps those coordinates to byte ranges.
01 / Separate genomics benchmark campaign
Large datasets. Measured results.
Selected ZCaps results from the historical genomics campaign. These measurements use a different system and build from the current ZMDC benchmark.
VCF chr21
1000 Genomes phase 3
99.26% reduction
134.6× compression ratio · Level 12
Targeted 4 MiB extraction: 0.16–0.17 s
Compare levels 1, 6 and 12
Level 1
- Data reduction
- 99.16%
- Compression ratio
- 119.5×
- Compression
- 1,958 MiB/s
- 4 MiB extraction
- 0.11 s
Level 6
- Data reduction
- 99.15%
- Compression ratio
- 117.6×
- Compression
- 1,876 MiB/s
- 4 MiB extraction
- 0.11 s
Level 12
- Data reduction
- 99.26%
- Compression ratio
- 134.6×
- Compression
- 1,488 MiB/s
- 4 MiB extraction
- 0.16–0.17 s
GTF
Gene annotation
98.63% reduction
73.07× compression ratio · Level 12
Compare levels 1, 6 and 12
Level 1
- Data reduction
- 97.39%
- Compression ratio
- 38.36×
- Compression
- 1,498 MiB/s
- 4 MiB extraction
- 0.15–0.16 s
Level 6
- Data reduction
- 98.21%
- Compression ratio
- 55.76×
- Compression
- 1,544 MiB/s
- 4 MiB extraction
- 0.16–0.18 s
Level 12
- Data reduction
- 98.63%
- Compression ratio
- 73.07×
- Compression
- 997 MiB/s
- 4 MiB extraction
- 0.22–0.25 s
SAM
Sequence alignment
87.37% reduction
7.92× compression ratio · Level 12
Compare levels 1, 6 and 12
Level 1
- Data reduction
- 82.73%
- Compression ratio
- 5.79×
- Compression
- 959 MiB/s
- 4 MiB extraction
- 0.27–0.29 s
Level 6
- Data reduction
- 85.27%
- Compression ratio
- 6.79×
- Compression
- 808 MiB/s
- 4 MiB extraction
- 0.28–0.33 s
Level 12
- Data reduction
- 87.37%
- Compression ratio
- 7.92×
- Compression
- 594 MiB/s
- 4 MiB extraction
- 0.42–0.46 s
FASTQ
Sequencing reads
77.68% reduction
4.48× compression ratio · Level 12
Compare levels 1, 6 and 12
Level 1
- Data reduction
- 74.16%
- Compression ratio
- 3.87×
- Compression
- 807 MiB/s
- 4 MiB extraction
- 0.35–0.37 s
Level 6
- Data reduction
- 76.19%
- Compression ratio
- 4.20×
- Compression
- 697 MiB/s
- 4 MiB extraction
- 0.39–0.42 s
Level 12
- Data reduction
- 77.68%
- Compression ratio
- 4.48×
- Compression
- 493 MiB/s
- 4 MiB extraction
- 0.63–0.67 s
GOA UniProt GAF
Functional annotation
94.63% reduction
18.63× compression ratio · Level 12
Compare levels 1, 6 and 12
Level 1
- Data reduction
- 92.26%
- Compression ratio
- 12.92×
- Compression
- 1,164 MiB/s
- 4 MiB extraction
- 0.302–0.315 s
Level 6
- Data reduction
- 93.85%
- Compression ratio
- 16.25×
- Compression
- 1,041 MiB/s
- 4 MiB extraction
- 0.301–0.338 s
Level 12
- Data reduction
- 94.63%
- Compression ratio
- 18.63×
- Compression
- 715 MiB/s
- 4 MiB extraction
- 0.437–0.488 s
0.11–0.75 smeasured targeted extraction envelope
30 formats · 3 compression levels · 90 compression runs · 270 targeted 4 MiB extraction runs · datasets up to 98.90 GiB.
SHA-256 validation of extracted ranges.
Benchmark system
Intel Core i9-13980HX
16 GiB RAM · NVMe SSD
Windows 11
Historical genomics campaign. Separate from the current ZMDC benchmark. Results are specific to the tested data, build and hardware.
Explore ZCaps technical evidence02 / Scientific data compression
Different structures. Adaptive compression.
Less data to store. The original data preserved. Selected Level 12 results from the historical genomics campaign.
| Dataset | Reduction | Ratio |
|---|---|---|
| VCF chr21 | 99.26% | 134.6× |
| GTF | 98.63% | 73.07× |
| Human FASTA | 97.28% | 36.78× |
| Human GFF3 | 97.23% | 36.08× |
| dbSNP JSON | 97.00% | 33.31× |
| OrthoXML | 96.56% | 29.06× |
| GOA GPA | 95.03% | 20.14× |
| GOA GAF | 94.63% | 18.63× |
| SAM | 87.37% | 7.92× |
| FASTQ | 77.68% | 4.48× |
Reduction = (1 − 1 / compression ratio) × 100, rounded to two decimals. Results compare ZCaps across datasets, not against competing compressors.
03 / Selective access at scale
Access the part you need.
Observed targeted extraction remained sub-second across the tested size range.
SCIENTIFIC ARCHIVE98.90 GiB example
Integration maps the selected region to a byte range ↓
Targeted 4 MiB extraction
Target the region. Decode the required chunks. Keep the rest compressed.
GFF31.65 GiB
JSON6.09 GiB
VCF11.23 GiB
VCF23.50 GiB
Pfam38.71 GiB
FASTA64.20 GiB
GOA GPA71.88 GiB
GOA GAF98.90 GiB
0.11–0.75 sObserved campaign envelope · targeted 4 MiB extraction
File sizes shown are documented examples. The extraction envelope describes the campaign, not individual latency values for each file or a universal latency guarantee.
04 / Bioinformatics infrastructure
One layer for scientific data workflows.
Genomic archives
Retain byte-exact scientific records with lossless bioinformatics compression and selective retrieval.
Variant datasets
Explore VCF compression for large variant collections while retaining targeted byte-range access.
Sequencing pipelines
Integrate FASTQ compression and SAM compression into workflows for reads and alignments.
Research repositories
Reduce repository storage while preserving original files for reproducible research.
Large annotation datasets
Compress GTF, GFF3 and functional annotation records using a data-adaptive engine.
Local scientific infrastructure
Keep compression and genomic random access inside infrastructure controlled by your organization.
High-volume bioinformatics storage
Use scientific data compression across large collections, processing independent files concurrently.
Discuss your workload