Personalizing medicine to an individual’s genetic profile

A new computational algorithm, applied to a large set of gene markers, has achieved greater accuracy than conventional methods in assessing individual risk for type 1 diabetes. A research team led by Dr Hakon Hakonarson of the Children’s Hospital of Philadelphia suggests that the technique, applied to appropriate complex multigenic diseases, should improve the prospects for personalizing medicine to an individual’s genetic profile – thus guiding prevention and treatment strategies. The study is published on 9 October in the open-access journal PLoS Genetics.

Genome-wide association studies (GWAS), in which automated genotyping tools scan the entire human genome seeking gene variants that contribute to disease risk, have yet to fulfil their potential in allowing physicians to accurately predict a person’s individual risk for a disease. For many of the recent studies using selected validated genes, the area under the curve (AUC), a method of measuring the accuracy of risk assessment, amounts to 0.55 to 0.60, little better than chance (0.50), and thus falling short of clinical usefulness.

Hakonarson’s team widened their search by looking at a broader collection of markers. By applying a "machine-learning" algorithm that finds interactions among data points, they were able to identify a large ensemble of genes that interact together. After applying this algorithm to a GWAS dataset for type 1 diabetes, the authors generated a model and validated it in two independent datasets. The model was accurate in separating type 1 diabetes cases from control subjects, achieving AUC scores in the mid-80s.

The researchers also stress the importance of choosing a target disease carefully. Type 1 diabetes is known to be highly heritable, with many risk-conferring genes concentrated in one region-the major histocompatibility complex. For other complex diseases that do not have major-effect genes in concentrated locations, this approach might not be as effective. Furthermore, the authors’ risk assessment model might not be applicable to mass population-level screening, but rather could be most useful in evaluating siblings of affected patients, who already are known to have a higher risk for the specific disease.

The authors believe their approach is more effective, and costs less, than human leukocyte antigen (HLA) testing, currently used to assess type 1 diabetes risk in clinical settings.

(Source: Public Library of Science: PLoS Genetics: October 2009)