The development of bioinformatics tools is an essential part of molecular biological research and the focus of our work. In order to empower experimental scientists, we have developed the MPI Bioinformatics Toolkit, which integrates state-of-the-art software for protein sequence and structure analysis into an intuitively usable platform . At this time our toolkit consists of 55 applications, most of which are interconnected, allowing job results to be forwarded between applications.
Around half of the tools available were developed in-house. One of our most important developments has been to harness pairwise hidden Markov model (HMM) comparisons for sequence analysis and structure prediction . With HHsearch, which is employed by our server HHpred  and by many others of our tools, we were the first to offer a fast and reliable algorithm for remote homology detection through HMM-HMM comparison. We also developed a protein classification tool, CLANS , which uses a variant of the Fruchterman-Reingold graph layout algorithm to analyze relationships in large sequence datasets. HHpred and CLANS are at the core of many of our papers [e.g. 5-8].
In addition to in-house developments, we have also integrated several public tools into our platform, some of which are offered as customized versions with extended functionality. We continuously strive to develop and integrate new tools into the toolkit and to improve its accessibility and user-friendliness.