Winner of the AAAI-07 Deployed Applications Awards!

Most microarray analysis software packages all do the same tasks, they find the most differentiated genes between two categories, make cluster dendrograms, etc, etc.

Some of them even allow you to use knowledge resources like the Gene Ontology to drill down deeper into the biological meaning of your analytical results.  

All this is valuable -- but ArrayGenius is different!

When microarray data is loaded into ArrayGenius, the first thing it does is compare your data to its internal ontologies -- built based on the Gene Ontology and proteomic databases -- and create an enhanced dataset.  The enhanced dataset tells you the degree to which various biological processes and protein families are expressed in your dataset.  ArrayGenius uses the enhanced dataset along with the original one, in all its subsequent analyses.

You can make cluster dendrograms -- both the familiar kind where you see which genes cluster together in the dataset, and a new kind, where you get to observe which biological processes cluster together in the dataset.

And then things get really interesting....

A common situation is where microarray data samples can be divided into two or more categories.  These categories may be Case vs. Control, they may represent different time points, etc.  In this sort of situation, you can ask ArrayGenius to learn classification models -- mathematical rules that predict whether a sample belongs to one category or another, using the gene expression values in the sample and also the inferred expression values of biological processes and protein families in the sample.  In many cases it finds extremely accurate rules -- our studies show that, after a bit of experimentation with parameter values, it generally beats the best algorithms from the academic literature.  Sometimes these rules are surprisingly simple, other times they're more complex.  

Generally ArrayGenius will learn a lot of classification models for a dataset, not just one or two -- and it can then study which genes, processes and families occur most often across all its models.  This is a novel way of detecting which genes, processes or families are most important to the biological phenomenon being studied in the dataset.  The most important features, in this sense, will often not be the ones most differentiated in expression between the categories of interest.  That's because ArrayGenius is figuring out which genes, processes and families are most important, not in terms of their solitary activity, but in terms of their interactions with other genes, processes and families.

And once you've gotten your results, you can interpret them via following hyperlinks into the Gene Ontology database, into PubMed, into various other online resources -- and into the BiomindDB, our own integrative data resource that provides useful functions like finding the research articles that focus on particular combinations of genes and processes.

ArrayGenius's novel approach provides a lot of information that more traditional microarray analysis doesn't:

• Ontological data integrated into the analytical process, so that classification rules and clusters involve biological processes and structures and families, not just raw gene expression values.  

• A sophisticated understanding of which biological processes are important to the phenomena under analysis

• Extremely accurate classification rules, useful for diagnostics and other purposes

New kinds of knowledge, indicating biological relationships and research directions that would otherwise go unnoticed.  A smarter way to analyze microarray data.

So you want to put the power of ArrayGenius to work on your data -- what's the next step?

ArrayGenius is an enterprise product, running on a small network of powerful Linux servers -- and with a 100% Web-based user interface, accessible through any Web browser.

If you want a dedicated ArrayGenius all for yourself, we also offer a "Managed Hosting" plan.  This means we host, in our server farm, a secure ArrayGenius installation that is entirely dedicated to your data, firewalled off from other machines.  This gives you total privacy and dedicated processing 24-7, 365 days a year.  Again: no software installation required.  Each individual in your organization can make an account, and log onto the site to upload or analyze data.  It's like having your own dedicated ArrayGenius OnDemand.

Finally, if you want to install ArrayGenius in-house, we offer two options: purchasing the software and installing it on your own Linux network, or purchasing a set of machines with Linux and ArrayGenius preinstalled.

If you're interested in managed hosting or an in-house installation, please email sales@biomind.com and one of our representatives will get back to you promptly.