SIMAP
SIMAP (Similarity Matrix of Proteins) was a BOINC-based volunteer computing project that maintained a large precalculated database of protein sequence similarities for use in bioinformatics research.[1][2] The database was freely accessible for scientific purposes and aimed to cover essentially all publicly known protein sequences.[3]
SIMAP was a joint project of the Technical University of Munich, the Helmholtz Zentrum München and the University of Vienna.[4] The project's BOINC-based computing arm, known as BOINCSIMAP, relied on volunteers donating idle CPU time to keep the similarity matrix up to date as new sequences were published.[3]
History
SIMAP was first described by Roland Arnold, Thomas Rattei and colleagues in 2005 as a database representing the protein similarity space, implemented with a calculation client capable of running both on Sun Gridengine clusters and as a BOINC application.[2] A more detailed description of the resulting database and its associated web services, including the ProtInfo query interface, was published the following year.[1]
The project initially received new work units at the beginning of each month. From 2010 onward, the inclusion of environmental metagenomic sequences into the database required much longer periods of continuous computation, sometimes lasting several months, with updates typically occurring twice a year.[4] In the fourth quarter of 2010, the project relocated from the Technical University of Munich to the University of Vienna after the failing electrical infrastructure at the Munich site made continued operation there impractical. Existing volunteers and users were required to detach from the old project URL and reattach to a new one at the University of Vienna.[4]
On 30 May 2014, project administrators announced that after a ten year history, SIMAP would leave the BOINC platform by the end of that year. The underlying SIMAP database and its research goals continued afterward using local hardware, described by the administrators as consisting of several hundred ordinary multi core CPUs running an SSE-optimised version of the Smith-Waterman algorithm.[4]
Methodology
SIMAP pre-calculated protein sequence similarities using the FASTA search heuristic, restricted by a static raw score threshold and the BLOSUM50 substitution matrix, expressed as
without limiting the maximum number of hits returned per sequence.[5] A separate application used hidden Markov models to search for protein domains within the same set of sequences.[4]
Because SIMAP represented an all-against-all similarity space, the computational cost of keeping the matrix current grew with the square of the number of contained sequences,
which is why the project relied on distributed volunteer computing rather than a single centralised system to keep pace with the growth of public sequence databases.[3] SIMAP took advantage of the approximate symmetry of alignment scores, treating the similarity of sequence A to sequence B as equal to that of B to A, which halved the amount of computation otherwise required and allowed the database to be updated incrementally as new sequences were added, without recomputing existing results.[5]
BOINC computing
On the BOINC platform, SIMAP work unit CPU times varied widely, ranging from about 15 minutes to 3 hours, with individual work units ranging from 1.5 to 2.2 megabytes in size, averaging around 2 MB.[4] Client software was provided that was optimised for SSE-enabled processors and x86-64 architectures; non-SSE applications were also available for older processors, though these required additional manual installation steps.[4] SIMAP supported Windows, Linux, macOS, Android and other Unix-like platforms.[4]
Because the database had, at various points, been brought up to date with all publicly known protein sequences and metagenomes, available work at any given time largely consisted of newly published sequences and metagenomes awaiting precomputation.[4]
Applications
The SIMAP database was used by individual researchers investigating the evolution and function of specific proteins, as well as by other bioinformatics projects that accessed SIMAP data directly through its public interfaces and web services, including integration with the PEDANT genome annotation system.[3][5] Use of the database was offered free of charge for education and public research.[6]
Publications
The following papers, listed on the official BOINC publications page, report scientific results that made direct use of SIMAP's BOINC-computed data.
- (2014).SIMAP—the database of all-against-all protein sequence similarities and annotations with new interfaces and increased coverage. Nucleic Acids Research. pp. D279–D284. DOI: 10.1093/nar/gkt970.
- Taylor, Simon J. E..(2011})."Distributed computing and modeling & simulation: Speeding up simulations and creating large models".In 2011 Winter Simulation Conference (WSC 2011).link.DOI: 10.1109/WSC.2011.6147748.
- (2010).SIMAP—a comprehensive database of pre-calculated protein sequence similarities, domains, annotations and clusters. Nucleic Acids Research. DOI: 10.1093/nar/gkp949.
- (2007).SIMAP structuring the network of protein similarities. Nucleic Acids Research. DOI: 10.1093/nar/gkm963.
- (2005).SIMAP--The similarity matrix of proteins. Bioinformatics. pp. ii42–ii46. DOI: 10.1093/bioinformatics/bti1107.
See also
External links
References
- ↑ 1.0 1.1 (2006).SIMAP: the similarity matrix of proteins. Nucleic Acids Research. pp. D252–D256. DOI: 10.1093/nar/gkj106.
- ↑ 2.0 2.1 (2005).SIMAP--The similarity matrix of proteins. Bioinformatics. pp. ii42–ii46. DOI: 10.1093/bioinformatics/bti1107.
- ↑ 3.0 3.1 3.2 3.3 Similarity Matrix of Proteins - SIMAP. BOINCSIMAP. Retrieved 2026-07-04.
- ↑ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 Similarity Matrix of Proteins. Wikipedia. Retrieved 2026-07-04.
- ↑ 5.0 5.1 5.2 SIMAP: the similarity matrix of proteins. PubMed Central. Retrieved 2026-07-04.
- ↑ SIMAP. BOINCSIMAP. Retrieved 2026-07-04.