BioWorld International Correspondent

ZICHRON YAAKOV, Israel - Compugen Ltd. is relying on medical chemistry basics in establishing its new subsidiary, Keddem Bioscience Ltd.

The independent company will focus on "a new design approach to small-molecule modulators for potentially any given protein target," said new co-CEO Dror Ofer, former vice president of Compugen's chemistry division, who added that the new endeavor will seek outside funding.

Arnon Levy, head of algorithm research in the former chemistry division and who also will co-manage, said that lead discovery, lead optimization and some ADME/Tox will be integrated into one process at Keddem, which is expected to produce small-molecule drug candidates.

The original program was initiated in 2000 as the chemistry division of Compugen, but now is being transferred to the new subsidiary.

Ofer, a Weizmann Institute of Science-trained physicist, told BioWorld International: "It's Keddem's objective to dramatically improve the success rate of the development of new drug products. Let's face it - the common technologies that rely on faster and bigger high-throughput screening of drug-like compounds mostly fail, and there is no reason to expect this to change in the absence of a theoretical basis."

Keddem, located in Ashkelon, Israel, sees difficult or intractable protein-protein interactions as relevant targets.

"Many of these are considered well validated but have so far not been dealt with, due to technical difficulties," Ofer said.

Keddem proposed creating a comprehensive set of less than 100,000 carefully designed molecules using a suite of proprietary algorithms to generate an alphabet of pharmacophore-binding site patterns. When synthesized, the compounds would be used in a screening process with a functional assay for the target of interest. The compendium of results would form a detailed 3-dimensional functional map of the binding site of the drug target.

No molecular modeling is involved in the discovery engine because the protein predicted in silico is functionally selected by the leads, i.e., each part of the map is directly confirmed in wet laboratory.

The hope is that the map will generate new, specific drug-like ligand designs by avoiding themes that have been mapped in other targets.

Ofer admits that Compugen's opinions are not widely held in big pharma. And he laments over the billions of dollars thrown away on a technology boom that in effect inhibits drug discovery, making it more costly to reveal an optimized lead chemical or preclinical drug candidate.

Millions of drug-like compounds are typically used in screening libraries, but, Ofer said, without a theoretically based search model, those numbers are insignificant.

Compugen's CEO Mor Amitai said about the new subsidiary, "Similar to our decision to establish Evogene as our ag-bio subsidiary, this enables Compugen to maintain its focus on predictive biology and on the creation and use of discovery engines for new therapeutic proteins and diagnostic markers."

Thus far, Compugen's discovery engines have uncovered six potential therapeutic proteins, including two novel splice variants of a previously known protein.