The Gene
Ontology is a powerful resource -- but it's not yet
complete. Many genes have not yet been assigned
functions.
Enter Biomind technology: we have devised
a clever combination of machine learning algorithms,
specifically designed to solve the problem of guessing
which GO categories a gene belongs to, based on analyzing
gene expression data across multiple experiments. We
call this software tool the OntoloGenius.
We have applied this technology to our
own database of public microarray datasets, thus making
provisional GO assignments for a host of human and yeast
genes. In May 2005 these assignments will be posted
online, as part of Biomind's mission to use our technology
to further bioscience in general. Current research
involves extending these results to the rat and mouse
genomes.
If you have a large number of proprietary
microarray datasets (say, 50+ datasets exploring the same
genes in different situations), and are curious what
your dataset implies about the functions of various genes,
then you may wish to apply OntoloGenius to your
data.
OntoloGenius is a command-line product
which runs in the Linux environment, and is easy to
integrate with all sorts of bioinformatic data processing
scripts.
Based on customer preference, we can also
apply OntoloGenius to customer data on a consulting basis
(see the Biomind
services page). Small customizations of
the OntoloGenius algorithms to your particular datasets
may yield even more powerful "gene function
inference" results.
If you're interested in applying the
OntoloGenius to your microarray data repository, e-mail info@biomind.com and we'll get back to you promptly.
Currently under development, the
Biomind PathwayGenius represents the next major
step in the realization of the Biomind vision: the use of
powerful AI technology to integrate diverse biological
knowledge to make novel scientific discoveries.
The center of the PathwayGenius is
the MindMap: a visual diagram based on a unique semantic network knowledge representation, integrating
information about biological relationships derived
from microarray data, protein-protein interaction data,
existing ontologies and pathway databases, and analysis of
biological texts.
Advanced AI inference
algorithms are applied to the relationships derived from
these various sources, and used to derive new
relationships, displayed in the MindMap -- creating a
visual repository of biological knowledge more
comprehensive, insightful and integrative than anything
constructed before.