Open for contributions

A number of small projects are opened at Mathbiol for small client side bioinformatics applications that take advantage of increasing computational capabilities of the browser. Typically, but not necessarily, these applications are designed to make the most of the server side functionalities of S3DB integrative infrastructure.
  1. PFGE Client - Pulse field Gel Electrophoresis.

Ongoing @ MathBiol:

  1. S3DB - semantic, distributed data service. The defining characteristic of this server-side DBMS, and of semantic web representations in general, is that the data model can be edited without compromising existing data. See our report in Nature Biotechnology for the motivation and overview of the architecture of this prototype application, which is made available to the public with open source.
  2. BiS - Bioinformatics Station - is a code distribution infrastructure that relies on MATLAB's runtime server. The purpose of BiS is to provide stand alone bioinformatics applications directly to the desktop of individual users without requiring installation or the acquisition of a software license.
  3. TCGA-GI - is developing a client graphic interfaces to integrate the visualization of the data produced by The Cancer Genome Atlas (TCGA)
    and that of the results of its analysis.
  4. S-Machine - is a software pipeline that manages time series data and its processing towards the identification of S-System models. This modeling formalism has the interesting property of being very generic while having parameters that can be mapped to a network topology. Therefore, the purpose of the S-Machine is the reverse engineering of Biological Networks from their time resolved manifestations.
  5. NRN - Nearest Related Neighbors - is a weighted nearest neighbor classifier designed to answer the question "have I seen this before?".
  6. ANN - Artificial Neural Networks - is a neural classifier toolset that is being continuously improved to include a variety of techniques for validation. More precisely, in addition to teh conventional cross-validation for early stopping, the toolbox now also provides variable addition by boosting and boostrapping of that addition with resampling of external validation datasets.