Sequence search · Software
CD-HIT
Clusters protein and nucleotide sequences by sequence identity to eliminate redundancy.
What CD-HIT is
CD-HIT is a sequence analysis program designed to cluster protein or nucleotide sequences by identity to remove redundancy from biological datasets. It was created by Weizhong Li and provides tools for standard clustering as well as comparisons between two separate sequence collections.
The software suite includes specialised programs such as cdhit for protein sequences, cdhit-est for nucleotide sequences, cdhit-2d and cdhit-est-2d for cross-dataset comparisons, cdhit-454 for sequencing reads, and psi-cd-hit for extended searches. It also includes utility scripts for converting, evaluating, filtering, and plotting cluster results, and is accessible both as command-line tools and as a web server.
How it works
CD-HIT partitions sequence databases into representative clusters based on sequence identity thresholds. Specialized binaries execute clustering for single datasets (such as cdhit and cdhit-est) or identify matches between two separate datasets (such as cdhit-2d and cdhit-est-2d). Additional utilities enable cluster renumbering, size distribution statistics, phylogenetic tree formatting, and export to XML or SQL table structures.
The software is written in C++ and Perl. It natively supports parallel execution via OpenMP multi-threading. Compressed sequence inputs in .gz format are handled directly via the zlib compression library.
Key facts
- Software version
- 4.8.1
- Primary application
- Sequence clustering and redundancy removal
- Input sequence types
- Protein and nucleotide sequences
- Supported input formats
- Uncompressed text and gzip-compressed (.gz) files
- Parallel processing
- OpenMP multi-threading
- External dependencies
- zlib (for .gz inputs), BLAST+ (for psi-cd-hit)
What you can use it for
- Protein redundancy reduction
- Cluster large collections of protein sequences using cdhit to produce non-redundant representative sequence sets.
- Nucleotide sequence clustering
- Group nucleotide datasets such as transcriptomes or gene collections using cdhit-est based on sequence identity.
- Cross-dataset sequence comparison
- Compare two distinct sequence files against one another using cdhit-2d or cdhit-est-2d to detect matching or unique entries.
- BLAST+-assisted clustering
- Run psi-cd-hit to perform sequence clustering using an external installation of BLAST+.
Run it here
CD-HIT is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
Strengths and limits
Good at
- Supports multi-threaded execution out of the box using OpenMP.
- Accepts gzip-compressed (.gz) sequence files directly, reducing storage overhead.
- Provides targeted tools for both single-dataset clustering and pairwise two-dataset sequence comparisons.
- Includes extensive Perl utilities for post-processing cluster outputs into formats like XML, trees, and database tables.
Watch out for
- Requires the zlib development library on the host system to enable reading compressed .gz input files.
- Requires a separate, locally installed BLAST+ executable in the system PATH to run psi-cd-hit.
Questions
Can I run CD-HIT online without installing it?
CD-HIT is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
What licence is CD-HIT under?
CD-HIT is published under its authors' own licence; the project page at https://github.com/weizhongli/cdhit states it.
Does CD-HIT read gzip-compressed sequence files?
Yes. Starting from version 4.8.1, CD-HIT supports .gz format input files directly, provided it is compiled with zlib library support.
Can CD-HIT run on multiple processor cores?
Yes. By default, CD-HIT compiles with multi-threading enabled via OpenMP, though it can optionally be built without OpenMP if needed.
Which version of BLAST is required for psi-cd-hit?
psi-cd-hit requires modern BLAST+ executables to be installed and available in the system PATH; legacy BLAST is not supported.
How can two different sequence collections be compared?
Two distinct collections can be compared against each other using cdhit-2d for proteins or cdhit-est-2d for nucleotide sequences.
Written from CD-HIT's own README and project page. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.