Title: Senior scientist, San Diego Supercomputing Center
Education: PhD, Rutgers University, 2000
Recommended by: Philip Bourne, University of California, San Diego
After spending several years at the frontlines of drug development at various pharmaceutical companies, Lei Xie decided to return to academia and do something about what he calls the "current crisis" in pharmaceutical research by changing the standard drug-discovery paradigm. To do this, Xie is developing dynamic-based algorithms for determining and comparing protein functional motifs, such as ligand binding pockets, antigen epitopes, protein-DNA/RNA binding sites, and allosteric pathways, on a structural proteome scale. In addition, he is working on designing accurate and efficient algorithms for predicting ligand-binding affinity by integrating structural bioinformatics, molecular modeling, and machine learning.
"One challenge is how to integrate all the 'omics data and put it all into a kind of holistic model so we can determine how the environmental perturbations, drugs, and other genetic alterations change the phenotype," he says. "Another area that remains difficult for us is predicting protein-ligand interactions. Those multiple interactions between the [proteins] and other molecules may have big impact on the phenotype, so how to predict and study these strong interactions on a genome scale is also a big area of interest."
Xie says that better molecular dynamic simulation software — although quite computationally intensive and difficult to apply on a genome scale — is a useful technique for studying a single molecule. "Metabolic modeling is also quite useful for studying systems, but you need to consider signal transduction and gene regulation, and these kinds of things are still difficult to incorporate all together," he says.
In the future, Xie would like to know the outcomes of inserting small molecule into a cell, in particular how certain small molecules of interest will interact with the whole proteome. "That would be very interesting as well as being able to take cancer genome or epigenomics data and put it all into a network or a context," he says. "Basically now most of the 'omics is only considered in a single way, but if you had a way to link it all together — how they are connected — that is quite important."
Publications of note
In 2009, Xie and Phil Bourne published a paper in PLoS Computational Biology in which they examine the origins of torcetrapib side effects. Torcetrapib is part of a new class of drug therapies for cardiovascular disease and it was withdrawn from phase III clinical trials when potentially deadly side effects appeared. The authors introduced a systematic strategy to study the human structural proteome and investigate the roles of off-targets effects on physiology and pathology using biochemical pathway analysis. Their findings suggest that the possible side effects of a new drug can be identified early in the development cycle and can even be minimized by fine-tuning multiple off-target interactions.
And the Nobel goes to ...
If Xie were to win the Nobel Prize, he would like it to be for changing the way people approach pharmacology from that of a single-interaction point of view to one that incorporates multiple interactions.