HISAT2 is a fast and sensitive alignment program for mapping next-generation sequencing reads (both DNA and RNA) to a population of human genomes (as well as to a single reference genome). Based on an extension of BWT for graphs [Sirén et al. 2014], we designed and implemented a graph FM index (GFM), an original approach and its first implementation to the best of our knowledge. In addition to using one global GFM index that represents a population of human genomes, HISAT2 uses a large set of small GFM indexes that collectively cover the whole genome (each index representing a genomic region of 56 Kbp, with 55,000 indexes needed to cover the human population). These small indexes (called local indexes), combined with several alignment strategies, enable rapid and accurate alignment of sequencing reads. This new indexing scheme is called a Hierarchical Graph FM index (HGFM).
Open Source Software

Site Map

News and Updates

    New releases and related tools will be announced through the Bowtie mailing list.

Getting Help

    Please use hisat2.genomics@gmail.com for private communications only. Please do not email technical questions to HISAT2 contributors directly.



H. sapiens, Ensembl GRCh38
genome 3.9 GB
genome_snp 4.6 GB
genome_tran 4.1 GB
genome_snp_tran 4.6 GB
H. sapiens, UCSC hg38
genome 4.1 GB
H. sapiens, UCSC hg38 and gene annotations referred to in the Nature Protocol paper
genome_tran 4.2 GB
H. sapiens, Ensembl GRCh37
genome 3.9 GB
genome_snp 4.5 GB
genome_tran 4.0 GB
genome_snp_tran 4.5 GB
H. sapiens, UCSC hg19
genome 3.9 GB
M. musculus, Ensembl GRCm38
genome 3.7 GB
genome_snp 4.0 GB
genome_tran 3.7 GB
genome_snp_tran 4.0 GB
M. musculus, UCSC mm10
genome 3.5 GB
R. norvegicus, UCSC rn6
genome 3.7 GB
D. melanogaster, Ensembl BDGP6
genome 0.2 GB
genome_tran 0.2 GB
D. melanogaster, UCSC dm6
genome 0.2 GB
C. elegans, Ensembl WBcel235
genome 0.1 GB
genome_tran 0.1 GB
C. elegans, UCSC ce10
genome 0.1 GB
S. cerevisiae, Ensembl R64-1-1
genome 0.01 GB
genome_tran 0.01 GB
S. cerevisiae, UCSC sacCer3
genome 0.01 GB

* genome: HFM index for reference
* genome_snp: HGFM index for reference plus SNPs
* genome_tran: HGFM index for reference plus transcripts
* genome_snp_tran: HGFM index for reference plus SNPs and transcripts
more indexes

Related Tools

  • HISAT: Fast and sensitive spliced alignment
  • Bowtie2: Ultrafast read alignment
  • TopHat2: Spliced read mapper for RNA-Seq
  • Cufflinks: Isoform assembly and quantitation for RNA-Seq
  • StringTie: Transcript assembly and quantification for RNA-Seq




HISAT, StringTie and Ballgown protocol published at Nature Protocols 8/11/2016

HISAT2 2.0.4 Windows binary available here, thanks to André Osório Falcão 5/24/2016

HISAT2 2.0.4 release 5/18/2016

Version 2.0.4 is a minor release with the following changes.
  • Improved template length estimation (the 9th column of the SAM format) of RNA-seq reads by taking introns into account.
  • Introduced two options, --remove-chrname and --add-chrname, to remove "chr" from reference names or add "chr" to reference names in the alignment output, respectively (the 3rd column of the SAM format).
  • Changed the maximum of mapping quality (the 5th column of the SAM format) from 255 to 60. Note that 255 is an undefined value according to the SAM manual and some programs would not work with this value (255) properly.
  • Fixed NH (number of hits) in the alignment output.
  • HISAT2 allows indels of any length pertaining to minimum alignment score (previously, the maximum length of indels was 3 bp).
  • Fixed several cases that alignment goes beyond reference sequences.
  • Fixed reporting duplicate alignments.

HISAT2 2.0.3-beta release 3/28/2016

Version 2.0.3-beta is a minor release with the following changes.
  • Fixed graph index building when using both SNPs and transcripts. As a result, genome_snp_tran indexes here on the HISAT2 website have been rebuilt.
  • Included some missing files needed to follow the small test example (see the manual for details).

HISAT2 2.0.2-beta release 3/17/2016

Note (3/19/2016): this version is slightly updated to handle reporting splice sites with the correct chromosome names.
Version 2.0.2-beta is a major release with the following changes.
  • Memory mappaped IO (--mm option) works now.
  • Building linear index can be now done using multi-threads.
  • Changed the minimum score for alignment in keeping with read lengths, so it's now --score-min L,0.0,-0.2, meaning a minimum score of -20 for 100-bp reads and -30 for 150-bp reads.
  • Fixed a bug that the same read was written into a file multiple times when --un-conc was used.
  • Fixed another bug that caused reads to map beyond reference sequences.
  • Introduced --haplotype option in the hisat2-build (index building), which is used with --snp option together to incorporate those SNP combinations present in the human population. This option also prevents graph construction from exploding due to exponential combinations of SNPs in small genomic regions.
  • Provided a new python script to extract SNPs and haplotypes from VCF files, hisat2_extract_snps_haplotypes_VCF.py
  • Changed several python script names as follows
    1. extract_splice_sites.py to hisat2_extract_splice_sites.py
    2. extract_exons.py to hisat2_extract_exons.py
    3. extract_snps.py to hisat2_extract_snps_haplotypes_UCSC.py

HISAT2 2.0.1-beta release 11/19/2015

Version 2.0.1-beta is a maintenance release with the following changes.
  • Fixed a bug that caused reads to map beyond reference sequences.
  • Fixed a deadlock issue that happened very rarely.
  • Fixed a bug that led to illegal memory access when reading SNP information.
  • Fixed a system-specific bug related to popcount instruction.

HISAT2 2.0.0-beta release 9/8/2015 - first release

We extended the BWT/FM index to incorporate genomic differences among individuals into the reference genome, while keeping memory requirements low enough to fit the entire index onto a desktop computer. Using this novel Hierarchical Graph FM index (HGFM) approach, we built a new alignment system, HISAT2, with an index that incorporates ~12.3M common SNPs from the dbSNP database. HISAT2 provides greater alignment accuracy for reads containing SNPs.
  • HISAT2's index size for the human reference genome and 12.3 million common SNPs is 6.2GB (the memory footprint of HISAT2 is 6.7GB). The SNPs consist of 11 million single nucleotide polymorphisms, 728,000 deletions, and 555,000 insertions. The insertions and deletions used in this index are small (usually <20bp).
  • HISAT2 comes with several index types:
    1. Hierarchical FM index (HFM) for a reference genome (index base: genome)
    2. Hierarchical Graph FM index (HGFM) for a reference genome plus SNPs (index base: genome_snp)
    3. Hierarchical Graph FM index (HGFM) for a reference genome plus transcripts (index base: genome_tran)
    4. Hierarchical Graph FM index (HGFM) for a reference genome plus SNPs and transcripts (index base: genome_snp_tran)
  • HISAT2 is a successor to both HISAT and TopHat2. We recommend that HISAT and TopHat2 users switch to HISAT2.
    1. HISAT2 can be considered an enhanced version of HISAT with many improvements and bug fixes. The alignment speed and memory requirements of HISAT2 are virtually the same as those of HISAT when using the HFM index (genome).
    2. When using graph-based indexes (HGFM), the runtime of HISAT2 is slightly slower than HISAT (30~80% additional CPU time).
    3. HISAT2 allows for mapping reads directly against transcripts, similar to that of TopHat2 (use genome_tran or genome_snp_tran).
  • When reads contain SNPs, the SNP information is provided as an optional field in the SAM output of HISAT2 (e.g., Zs:Z:1|S|rs3747203,97|S|rs16990981 - see the manual for details). This feature enables fast and sensitive genotyping in downstream analyses. Note that there is no alignment penalty for mismatches, insertions, and deletions if they correspond to known SNPs.
  • HISAT2 provides options for transcript assemblers (e.g., StringTie and Cufflinks) to work better with the alignment from HISAT2 (see options such as --dta and --dta-cufflinks).
  • Some slides about HISAT2 are found here and we are preparing detailed documention.
  • We plan to incorporate a larger set of SNPs and structural variations (SV) into this index (e.g., long insertions/deletions, inversions, and translocations).

The HISAT2 source code is available in a public GitHub repository (5/30/2015).