zzh
/
BWA-FastAlign
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Rename FastAlign to BWA-FastAlign in README

This commit is contained in:
Yewen Li 2026-01-04 00:09:48 +08:00 committed by GitHub
parent 643832af09
commit 98f8a27a7c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 9 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
README.md
View File

@ -1,6 +1,6 @@
# FastAlign: Faster and Cheaper Sequence Alignment on Commercial CPUs
# BWA-FastAlign: Faster and Cheaper Sequence Alignment on Commercial CPUs
**FastAlign** is a high-performance, cost-efficient software package for mapping low-divergent sequences against a large reference genome, such as the human genome.
**BWA-FastAlign** is a high-performance, cost-efficient software package for mapping low-divergent sequences against a large reference genome, such as the human genome.
It is designed as a drop-in replacement for the de facto standard **BWA-MEM**, offering **2.27× ∼ 3.28× throughput speedup** and **2.54× ∼ 5.65× cost reductions** on standard CPU servers, while guaranteeing **100% identical output** (SAM/BAM) to BWA-MEM.
@ -14,7 +14,7 @@ It is designed as a drop-in replacement for the de facto standard **BWA-MEM**, o
## 🔧 Technical Innovations
FastAlign revitalizes the traditional alignment pipeline with two core algorithmic contributions:
BWA-FastAlign revitalizes the traditional alignment pipeline with two core algorithmic contributions:
1. **Multi-Stage Seeding (Hybrid Index)**
* Combines **Kmer-Index**, **FMT-Index** (Enhanced FM-Index with prefetching), and **Direct-Index**.
@ -22,7 +22,7 @@ FastAlign revitalizes the traditional alignment pipeline with two core algorithm
* Achieves an **18.92× improvement in memory efficiency** (bases processed per GB per second).
2. **Intra-Query Parallel Seed-Extension**
* Unlike BWA-MEM2 (which uses inter-query parallelism and suffers from load imbalance), FastAlign parallelizes the Smith-Waterman alignment *within* a single query.
* Unlike BWA-MEM2 (which uses inter-query parallelism and suffers from load imbalance), BWA-FastAlign parallelizes the Smith-Waterman alignment *within* a single query.
* Includes **Dynamic Pruning** to skip zero-alignment scores.
* Implements a **Sliding Window** mechanism to reduce costly memory gather operations.
* Achieves **3.45× higher SIMD utilization**, performing consistently well on both WGS (Whole Genome Sequencing) and WES (Whole Exome Sequencing) data.
@ -37,13 +37,13 @@ FastAlign revitalizes the traditional alignment pipeline with two core algorithm
### Compilation
```bash
git clone https://github.com/your-username/FastAlign.git
cd FastAlign
git clone https://github.com/your-username/BWA-FastAlign.git
cd BWA-FastAlign
make
```
## 📖 Usage
FastAlign follows the same command-line interface as BWA-MEM.
BWA-FastAlign follows the same command-line interface as BWA-MEM.
1. **Index the Reference.** Before alignment, you must index your reference genome (e.g., human_g1k_v37.fasta).
```bash
@ -63,10 +63,10 @@ FastAlign follows the same command-line interface as BWA-MEM.
./fastalign mem -t 64 ref.fa read1.fq read2.fq > aln.sam
```
3. **Options.** FastAlign supports the standard BWA-MEM options. Run ./fastalign mem to see the full list.
3. **Options.** BWA-FastAlign supports the standard BWA-MEM options. Run ./fastalign mem to see the full list.
## 📜 Citation
If you find FastAlign is useful in your research, please cite our paper:
If you find BWA-FastAlign is useful in your research, please cite our paper:
```bibtex
@inproceedings{fastalign2026,
title={Faster and Cheaper: Pushing the Sequence Alignment Throughput with Commercial CPUs},