Ultra-fast alignment of large-scale DNA and protein sequences

MUMmer is a system for extremely fast alignment of large-scale DNA and protein sequences. For example, MUMmer v 2.1 can align two completely bacterial genomes of 3-4 million base pairs (Mbp) each in under 30 seconds. MUMmer was developed by The Institute for Genomic Research (S.L. Salzberg and A. Phillippy) and Celera Genomics (A.L. Delcher).

MUMmer can also align incomplete genomes; it handles the 100s or 1000s of contigs from a shotgun sequencing project with ease, and will align them to another set of contigs or a genome, using the NUCmer utility included with the system. The PROmer utility takes this a step further by generating alignments based upon the six-frame translations of both input sequences. PROmer permits the alignment of genomes for which the proteins are similar but the DNA sequence is too divergent to detect similarity.

MUMmer's namesake program originally stood for M aximal U nique M atcher, however in subsequent versions the meaning of unique has been skewed. MUMmer achieves its high performance by using a very efficient data structure known as a suffix tree. This data structure can be both constructed and searched in linear time, making it ideal for large scale pattern matching. To save memory, only the reference sequence(s) is used to construct the suffix tree and the query sequences are then streamed through the data structure while all of the maximal exact matches are extracted and displayed to the user.

We are currently working on a CGI interface for MUMmer 3.19 to make it easy to compare a wide variety of genomes. TIGR has their own version of a web interface, although it is limited by a their own proprietary list.