Researchers from big pharma, academia, and proteomics equipment suppliers at last week’s ACI (Active Communications International) proteomics meeting in Boston, chaired by Steven Naylor, chief technology officer at Beyond Genomics, got a look at a broad sampling of the various approaches to applying proteomics to drug discovery. In fact, the only thing missing was a talk from a HUPO representative touting the advantages of large-scale proteomics projects. In its stead, attendees learned about new developments in peptide mass spectrometry software, quantitative protein labeling technology, and advances in liquid chromatography-based approaches for protein separation, among other topics. Highlights from the conference — ACI’s first effort in the area of proteomics — follow:
Using DTASelect to Augment SEQUEST
Hayes McDonald, a senior research associate in John Yates’ lab at the Scripps Research Institute in La Jolla, spoke on his group’s application of “shotgun” proteomics to identify the components of protein mixtures, and described a new program, DTASelect, which ameliorates the analysis of peptide mass spectrometry data when using the SEQUEST program.
For running protein identification experiments using online liquid chromatography, the DTASelect algorithm applies confidence filters to SEQUEST output data, increasing the likelihood that the software is accurately predicting which proteins are present in the sample, McDonald said. While Yates’ group had this capability before DTASelect, the old programs did not allow the user to modify specific confidence filters, and were awkward because users had to run several different programs sequentially, he added.
“For certain matches you have more flexibility [with DTASelect] in applying confidence filters; in fact in the default filterings between 95 and 100 percent of the hits are real,” McDonald said. “It really reduces the amount of work you have in going through data from a MudPIT run. [But] if you want to do more work, you can modify the filters and try to mine the data for the maximum amount of information in it.”
Dave Tabb, the member of Yates’ group who wrote the program, said that 13 or 14 different academic groups are using DTASelect, which is available free-of-charge for academic users, and that several licensing agreements with companies are “in progress.” Tabb said he wasn’t sure whether Thermo Finnigan, which acquired the license to distribute SEQUEST from the University of Washington, would try to license the rights to distribute DTASelect as well. “I’m not so sure whether their emphasis would be to create something in-house that does the same things that DTASelect does, or whether they’ll want to license DTASelect itself,” Tabb said. “But yes we’re interested in getting them involved in the project.”
Hefta Brings Hefty Resources to Bear on Proteomics
Stan Hefta, executive director of proteomics at Bristol-Myers Squibb, spoke on his group’s approach to applying proteomics in the context of big pharma. Hefta, who began working in proteomics at BMS in 1997 as part of the applied genomics department, soon spun out his operation into a separate proteomics department, and now assists the company in validating protein targets, identifying posttranslational modifications, and discovering biomarkers for use in clinical trials.
Hefta has built his program around 2D gel technology for separating proteins, but has also expanded into liquid chromatography-based approaches using ion exchange and reverse phase columns, he said. His group relies on MALDI-TOFs, as well as Q-TOF and Thermo Finnigan LCQ ion trap mass spectrometers for performing MS/MS analyses of his protein samples.
Hefta has made a specialty of mass spectrometry data processing, having directed the group’s in-house software development and hardware acquisitions. His operation now employs a Linux cluster with several hundred CPUs, he said, and can currently process between 100,000 and 150,000 mass spectra per day, he added.
As an example of his group’s capabilities, Hefta described a project to identify the intermediates involved in triggering the differentiation of osteoclasts, the cells responsible for osteoporosis-related bone decomposition. When 2D gel analysis failed to pick out the low-abundance kinases involved in the mechanism, he applied immobilized metal-affinity chromatography (IMAC) to identify the intermediates, developed phosphospecific antibodies to isolate them, and helped BMS’ chemistry group synthesize small molecule inhibitors against them.
During the Q&A, Hefta disclosed that the anti-infective therapeutic program at BMS had been downgraded, but that developing antiviral drugs is still a priority for the company.
Amersham to Commericialize James’ Labeling Reagents
Peter James, a professor in the departent of electrical measurements at Lund University in Sweden, described his efforts to develop both 2D gel and liquid chromatography-based methods for separating proteins, as well as his patent-pending technique for quantifying protein expression using N-terminal labels.
James’ gel platform relies primarily on 2D DIGE (differential in-gel electrophoresis), a technology marketed by Amersham Biosciences, which also has rights to commercialize James’ labeling chemistry. The labels, which differ according to the number of heavy and light isotopes of hydrogen, allow James to quantify differences in protein expression between two protein samples as part of the DIGE experiment.
James said the labeling approach to 2D gel analysis avoids one typical disadvantage of working with gels — that the spots contain more than one protein — because the labels help resolve the proteins during MALDI-TOF analysis. “I don’t care how many proteins are in that damn spot,” he said. The same labels are also applicable to quantitative analysis of proteins using liquid chromatography, as well as analyzing posttranslationally-modified proteins, he said.
Amersham is currently working to commercialize the labels, and James estimated that the company would need a year and a half to bring the reagents to market.