DNA2.0 Inc. Co-Founders Reports Technology For Intelligent Protein Engineering

Menlo Park, CA
Friday, August 15, 2003

DNA2.0 Inc. announced today the publication of a scientific review entitled 'Putting engineering back into protein engineering" (Gustafsson et al., Current Opinion in Biotechnology 14 (4):366-370). The DNA2.0 co-founders report the emergence of modern heuristic statistics as applied to protein engineering. Modern heuristic statistics permits protein engineering by synthesizing information-rich genes and maximizing the capture of sequence-function correlation. This improved technology leads to rapid and cost efficient sequence modifications of proteins to produce new and functionally improved protein variants.

Protein engineering has classically been approached from two diametrically opposed directions: Rational design and Directed evolution. Rational design attempts to understand protein structure and function at a complete mechanistic level so that any desired change can be effected by calculations of first principle. Directed evolution attempts to find a desired solution by empirically testing many different variants. Modern heuristic statistics applied to protein engineering is a synthesis of empirical data and a rational analysis of that information. The heuristic statistics technology is well established in a multitude of different industries that face the same challenges of high system complexity and abundance of variables. Such industries include small molecule QSAR, optimization of manufacturing and telecom. The application in protein engineering has until now largely been limited by the prohibitive cost of gene synthesis. As cost of gene synthesis is coming down, this alternative protein engineering approach is primed to take off.

The current state-of-the-art of heuristic statistics for protein engineering is described in the publication, and a number of examples including Neural Nets, regression based algorithms and D-Optimal design are discussed in the context of optimizing the engineering algorithm.

By this relatively straight forward approach, any protein can be converted into an improved variant, even for functionality that is difficult or impossible to replicate in high throughput assays. This is due to the substantially lower number of variants needed to be tested as compared to directed evolution, and substantially higher hit rate as compared to rational design.

DNA2.0 Inc. has developed a suit of algorithms, software, and an in-house gene synthesis process to fully take advantage of this protein engineering paradigm shift. DNA2.0 President Dr. Jeremy Minshull said: "Modern heuristic statistics provides an opportunity to optimize any protein in any functional direction much more efficiently than has previously been attainable. DNA2.0 is now applying the technology in conjunction with our low cost/high speed gene synthesis to put engineering back into protein engineering."



DNA2.0 Inc. is developing a platform technology to rapidly and cost efficiently produce statistically optimally distributed gene variants. These gene variants are synthesized individually and are tested for functional activity. The data is integrated and mapped as a mega dimensional projection of the correlated sequencefunction space. The DNA2.0 Inc. technology is uniquely suited to bridge the information gap created by the recent explosion of genomic sequence information.


For more information, please contact:
DNA2.0 Corporate Communications