Report on an In-House Seminar on Genomic Medicine

BULLETIN FOR MEDICAL PRACTITIONERS

Dr Audrey Tan, Registrar, National Skin Centre

Introduction
The in-house seminar on Genomic Medicine was held on 31 August 2002. The 4 distinguished speakers from the Genome Institute of Singapore (GIS) discussed the following topics:

I. Introduction to Genomic Medicine in Singapore by Prof Edison Liu, Director of GIS.

II. Designing peptide vaccines by A/Prof Ren Ee Chee, Deputy Director of GIS.

III. Current approaches to mapping and identifying disease-predisposing genetic variation by Dr Mark Seielstad, Group Leader, Department of Population Genetics.

IV. Seeking needles in haystacks: proteomics approaches for disease marker discovery by Dr Eastwood Leung, Group Leader, Department of Proteomics.

I. Introduction to Genomic Medicine in Singapore
The goal of genomics is to uncover the structure and sequence of the genetic material that defines an organism. The strategic goals of the Genome Institute of Singapore include the following:

• to promote genetic research and to establish a genomics infrastructure in Singapore
• to provide international visibility for Singapore's life sciences programmes
• to support training of manpower in life sciences for Singapore and to attract scientific talent to Singapore

Singapore is a unique population laboratory due to its compact size, excellent medical facilities, research focus, research-friendly population and ability to develop a robust disease registry.

II. Designing peptide vaccines
Conventional vaccines comprise mainly a single attenuated strain or a single subunit polypeptide. Such vaccines are used worldwide and often elicit variable responses and at times morbidity and eventual mortality in a small number of individuals. Another consistent limitation of such vaccines is that it cannot cope with escape variants, as clearly demonstrated in the HIV and hepatitis C scenario.

A series of immunology-bioinformatics-proteomics work process by which peptide vaccines in silico can be designed have been created. This is accomplished by first generating a series of recombinant HLA molecules in bacteria and purifying the inclusion bodies in bulk to produce purified HLA protein. Together with beta-2 microglobulin and a known synthetic peptide, the peptide-MHC complex is refolded in the presence of denaturing urea conditions. Data on the quality of binding is then collected and fed into a computational algorithm as training sets. With a large number of data points, the algorithm will "learn" and eventually allow prediction of how antigenic peptides are to be fitted into an individual's HLA molecule accurately.

In future, the way to design vaccines will be to download a particular peptide sequence from the published genome of an infectious agent and process it by informatics and modeling protocols to yield a shortlist of candidate vaccine peptides. These will then be validated using functional and in vivo assays to select the best peptides. By constructing a combination peptide vaccine, it will be possible to provide a vaccine that is optimal for antigen presentation and eliciting a strong immune response across a diverse group of HLA makeup.

III. Current approaches to mapping and identifying disease predisposing genetic variation
The classic approach to mapping and identifying gene loci of disease manifestations is the family-based linkage analysis. This method works well for simple Mendelian diseases.

Complex and chronic diseases are frequently presumed to have a significant genetic contribution to their etiology. While often correct, the task of identifying such genetic variation has been difficult in practice, with few successes to date. Much of the difficulty lies in the complex, multifactorial nature of these diseases, with each gene exerting only a minor effect on the disease process.

As a result, genetic studies will require large clinical samples of patients and controls in order to have sufficient power to detect the relevant genetic loci. This requires increasingly close collaboration between practising clinicians and bench scientists. In his lecture Dr. Seielstad highlighted several methodological approaches to disease gene mapping, with discussion of their strengths and weaknesses.

IV. Seeking needles in haystacks: proteomics approaches for disease marker discovery
The process of disease marker discovery by using proteomics technologies is analogous to seeking needles in haystacks. There are basically two approaches:

1. Maximization of separation of complex samples such as body fluids so that biomarkers can be observed without the hindering of major proteins.
2. Specific affinity capture of molecules of interest with removal of non-specific molecules.

The first approach is exemplified by two-dimensional gel electrophoresis or multi-dimensional liquid chromatography. Proteins are separated based on their charge, molecular weight, or hydrophobicity. Proteins of interest are revealed after comparison of protein spot patterns or peak profiles between normal and diseased samples. Mass spectrometry identifies differential proteins. The second approach of affinity capture is exemplified by SELDI® (surface-enhanced laser desorption and ionization) technology. Proteins bind onto chip surfaces based on charge, hydrophobicity, or metal chelating affinity. Captured proteins are resolved by time-of-flight mass spectrometry. These two complementary approaches provide the maximal coverage of protein biomarker discovery process.

DEDICATED TO EXCELLENCE IN DERMATOLOGY
By National Skin Centre (Singapore)
Copyright (C) 1995 - National Skin Centre (Singapore)

top^