Abstracts

Human Genome Diversity

Variation of most quantitative traits is assumed to be determined by many genes interact-ing with each other and an environment. Quantitative trait loci (QTL) are usually being identified by means of positional (genetic) mapping based on statistical analysis. Tradi-tional statistical methods of genetic mapping, which made a major impact into identifica-tion of genes of Mendelian disorders, were found to be inadequate for QTL mapping . For a long time the development of new methods of statistical analysis was considered as a clue to further progress in identification of QTLs. In the last decade, a variety of new approaches to QTL mapping have been developed. These methods relied mainly on an increased computational efficiency of realization of well-known (parametric and non-parametric methods of linkage analysis) and novel methods (multipoint mapping, vari-ance components analysis, combined linkage/association studies) (Liu, 1998). The devel-opment of new powerful and robust methods has led to a certain progress in QTL-mapping: the number of studies in which linkage of QTLs with genetic markers was de-tected is growing every year (Cargill et al., 1999).

At the same time far less progress in identification of QTLs has been achieved. The results of QTL-mapping are often contradictory, e.g. close linkage found in one data set is rarely de-tected in other set (Ostander & Stanford, 2000). In some cases linkage analysis indicates re-gion of putative gene location where no structural gene is know. This makes impossible to choose a candidate gene. These findings may lead to conclusion that most of variation of quantitative traits is due to variation in genes responsible for regulation of expression of puta-tive gene, but not to variation in the structural gene itself (Weiss & Terwilliger, 2000). Mod-ern state of science allows to predict consequences of a mutation in a coding DNA sequences (gene - protein - function - trait), however, our knowledge does not allow to predict conse-quences of a mutation in a uncoding (regulatory) DNA sequences. It turns out that statistical methods, developed in last decade, are sufficient of QTL-mapping for a given data set, but do not help much in QTL- identification. The problem of QTL-identification appears to be a biological, rather than a statistical problem. The real questions are: what are the biological mechanisms of gene-regulation and how does the polymorphism of regulatory genes affect quantitative traits.

Cargill M et al. Nat Genet, 1999, 22:231-238
Liu BH Statistical genomics: linkage, mapping and QTL analysis, CRC Press LLC, 1998
Ostander EA, Stanford JL Am J Hum Genet, 2000,67:1367-1375
Weiss KM, Terwilliger JD Nat Genet, 2000, 26:151-157

Note. Abstracts are published in author's edition

|Home Page| |English Part|

Go to Home