The MEME/MAST project allows biomedical scientists to model molecular sequences and identify relationships and patterns among them. This aids in understanding the structure and function of genes and proteins in the cell. MEME is a pattern discovery tool, based on statistical learning algorithms, that finds sequence patterns in groups of proteins or genes. These patterns are useful for understanding the important biological features of the sequences. Furthermore, they can be used by the MAST algorithm to identify genes and proteins that share the patterns found by MEME, and are therefore likely to be functionally or evolutionarily related to the original group of sequences. MEME and MAST comprise an extremely powerful method for identifying distant but important relationships among biological molecules. NBCR provides MEME and MAST via """"""""transparent supercomputing"""""""" to the biological research community, via an easy-to-use world wide web interface. We have continued to improve both MEME and MAST. MAST has been enhanced with colored, graphical output to make distant relationships easy to visualize, and the ability to search DNA (genes) directly for patterns discovered in protein sequences using MEME. We have added new capabilities to MEME including the ability to compensate for genetic sequences that are over-represented in the training data, and to find patterns that better discriminate between closely related groups of molecules. In addition, we have improved MEME's output to allow it to interface better with other sequence analysis tools. For example, MEME motifs can be viewed as LOGOS, used to search the BLOCKS database of protein motifs, and used to build phylogenetic trees via the tools provided at the BLOCKS website (www.blocks.fhcrc.org/blocks/process_blocks.html). Scientists are submitting approximately 300 MEME runs and 180 MAST runs to the supercomputers at SDSC each month (Fig.1). The enhanced capabilities and improved user interfaces of the programs will doubtless attract more users when we introduce new versions (MEME 2.2; MAST 2.2) this month. About 1000 users have run MEME queries via the SDSC server, and about 20 users have run more than 20 MEME queries. Approximately 500 users have run MAST queries via the SDSC server. Figure 1: MEME usage on the Cray T3E and MAST usage on the DEC Alpha ``farm'' at SDSC via the SDSC MEME/MAST web site.
Showing the most recent 10 out of 270 publications
