• Download a PDF version of this case study.

Discovery Machine forges new pathways in gene mapping, pharmaceuticals and biocomputing. Data fusion capabilities will aid homeland defense.

Scientists now have the ability to build working models, simulations and intelligent agents without the need for programming.

Breaking the computational logjam in gene mapping research

A major drive in medical research today is the intensive search for genes that affect susceptibility to disease. Yet scientists worldwide face the same obstacle: analyzing, managing and tracking data at a rate that keeps up with its generation—and with the new analysis techniques for processing it. Serious analysis bottlenecks are being caused by computational backlogs. Breaking through the computational logjam will revolutionize medical research.

The Challenge

The analysis of diverse and disparate data—which requires advanced pattern recognition and data fusion techniques—has become a 21st century imperative. Data fusion itself—the seamless integration of data from widely varying formats— represents a substantial obstacle.

Among the agencies, institutions and corporations throughout the world searching for practical solutions are the National Science Foundation, GlaxoSmithKline—one of the world’s foremost pharmaceutical companies—and North Carolina State University, a leading research institution.

Awarded a grant by the National Science Foundation’s Small Business Innovative Research (SBIR) program, Discovery Machine Inc. worked with GlaxoSmithKline’s Department of Population Genetics and NC State’s renowned Bioinformatics Research Center to model complex genetic data.

The endeavor involved computer analysis of a vast amount of genetic data for thousands of individuals. Maintaining and updating the computer scripts and parameters used to run complex analysis programs is a significant challenge. So is managing and tracking the results.

The Key Solution

Working directly with North Carolina State University geneticists, Discovery Machine enabled the researchers themselves to build their research expertise into a gene-mapping model without the need for programming. The model also allowed a novice analyst to immediately run a set of unfiltered, inconsistent pedigree data. Connections were made to a set of database integration features that were specified, designed and implemented to meet GlaxoSmithKline requirements. Also included was the ability to connect to rapidly changing analysis routines, needed by pharmaceutical and academic research laboratories alike.

Importantly, Discovery Machine’s model ensured that earlier versions of programs remain valid and can be automatically integrated into new or existing databases. The Discovery Machine model is rapidly scalable and can seamlessly incorporate new statistical algorithms.

The Result

This solution increases the speed and efficiency of scientific research, and it frees up scientists to focus upon core research concerns. The abilities to automatically incorporate new analysis programs, parameters and data and to re-evaluate past data open up entire new avenues of investigation.

The Discovery Machine solution is particularly valuable because knowledge of a research laboratory’s unique computer scripts can be lost when key personnel change jobs or retire. The loss of personnel and their knowledge also can block re-evaluations of past data, because the original analysis parameters are lost. System upgrades also can change parameters.

© 2005-2007 Discovery Machine, Inc. All rights Reserved. | 454 Pine Street | Williamsport, PA 17701