fcount-rs¶
Ultrafast feature counting for RNA-seq data. A high-performance Rust alternative to featureCounts.
Performance¶
fcount-rs is 2× faster than featureCounts with 8 threads on typical RNA-seq data:
| Tool | Threads | Time (s) | Speedup |
|---|---|---|---|
| featureCounts | 1 | 3.34 | — |
| featureCounts | 8 | 1.45 | — |
| fcount-rs | 1 | 2.26 | 1.5× |
| fcount-rs | 8 | 0.71 | 2.0× |
Benchmark: chr22 subset (~1.4M paired-end reads). featureCounts v2.1.1.
Features¶
- Fast: 2× faster than featureCounts with parallel processing, custom GTF parser, and efficient data structures
- Pure Rust: No C dependencies, uses the noodles library for BAM/SAM parsing
- Compatible: Drop-in replacement for featureCounts with similar CLI interface
- Flexible: Supports single-end and paired-end data, stranded protocols, and multi-mapping reads
Quick Start¶
# Install
cargo install fcount-rs
# Count reads
fcount-rs -a annotation.gtf -o counts.txt sample.bam