Name:
Eric John Martin
Country:
USA
City:
El Cerrito
State/Province:
CA
ACS Member:
Member
Local Section:
L601,California
International Chapter:
Technical Division Membership:
D507,Chemical Information Division;D510,Computers in Chemistry Division
Technical Division Membership Emeritus:
Subdivision Member:
ACS Activities:
Company/Institution Name:
Career Stage:
Professional
Research and Special Interests:
Current Responsibilities: Manage Computational Chemistry group. Cost center manager. Developing new computational methods for kinase-family drug discovery. Interactive structure-based drug design for therapeutic projects, most recently Eg5.
Notable scientific contributions:
- GRASP (1988, while at Dow, in collaboration with Greg McRae, Carnegie Mellon) First method and program to predict and analyze the likelihood of groundwater contamination by a pesticide, across wide geographic areas and usage practices. It combines a leaching model, a soils database, and a stochastic weather model, using global sensitivity analysis. Now used routinely by the EPA to evaluate pesticide safety.
- MAKESPACE/TAILOR (1993) First method and program for combinatorial library design: maximizing library diversity while focusing it toward specific targets and druggable properties. Combinatorial library design is now widely practiced throughout the drug industry.
- MAGNET (with Hanneke Jansen, 1999) First method and program to customize docking functions to a particular target by adding weighted interaction features that compensate for errors in general docking functions. Several commercial, academic, and proprietary docking programs now include Magnet-like target customization.
- AUTOSHIM (with David Sullivan, 2005) An extension of MAGNET to produce quantitative docking models by simultaneously training interaction feature weights ("shims"), and pose selection against assay data. This dramatically improves predictions of affinity over conventional docking.
- "SURROGATE AUTOSHIM" (with David Sullivan, 2006) A much more predictive alternative to homology model docking. It extends AUTOSHIM by building quantitative docking models for members of a protein family (i.e. kinases) which have no crystal structure, by "shimming" a universal surrogate ensemble of crystal structures for several other diverse family members to reproduce assay data for new targets. The Novartis archive was pre-docked into a universal kinase surrogate, so kinase docking is now as fast as QSAR.
Primary Current Research Interests:
- Interactive structure-based drug design.
- "AUTOSHIM": i.e. creating target-customized docking functions by generating interaction features for an ensemble of crystal structures, parameterizing regression models on the features, and iteratively refining the pose selection with the parameterization.
- Kinase target family modeling:
1. "Chemo-genometrics", i.e. the use of historical activity and structural information against a panel of kinases to predict activity profiles against new kinases.
2. "Surrogate docking", i.e. "autoshimming" an ensemble of structures for known kinase to predict the binding of a new kinase with no crystal structure.
3. Chemogenomic-based selection of "drugable" kinase target profiles.
- High-throughput computing on Beowulf clusters.
- Conformational analysis and active-analog approaches for drug discovery using ab initio quantum methods.
- Targeted, structure-based and property-tailored combinatorial library design.
All of my approaches explicitly aim to augment molecular modeling and statistical methods with empirical data and with the intuition and experience of practicing medicinal chemists.
Additional Research Interests: Methods for calculating and maximizing the diversity of combinatorial libraries for drug screening, and for biasing diversity libraries toward physical and chemical properties required for
Area of Expertise:
Computational Chemistry: Kinase-Family modeling, Iterative screening, Hit-finding technologies, Target-customized scoring functions, combinatorial library design, structure-based drug design, methods development, analytical instruments to support compuational chemistry.
Years of Expertise:
1-5