CiteULike: neteler's Schwartz

Landscape genetics: combining landscape ecology and population genetics

Stephanie Manel — 2007-10-09T09:54:02-00:00

Trends in Ecology & Evolution, Vol. 18, No. 4. (April 2003), pp. 189-197.

Understanding the processes and patterns of gene flow and local adaptation requires a detailed knowledge of how landscape characteristics structure populations. This understanding is crucial, not only for improving ecological knowledge, but also for managing properly the genetic diversity of threatened and endangered populations. For nearly 80 years, population geneticists have investigated how physiognomy and other landscape features have influenced genetic variation within and between populations. They have relied on sampling populations that have been identified beforehand because most population genetics methods have required discrete populations. However, a new approach has emerged for analyzing spatial genetic data without requiring that discrete populations be identified in advance. This approach, landscape genetics, promises to facilitate our understanding of how geographical and environmental features structure genetic variation at both the population and individual levels, and has implications for ecology, evolution and conservation biology. It differs from other genetic approaches, such as phylogeography, in that it tends to focus on processes at finer spatial and temporal scales. Here, we discuss, from a population genetic perspective, the current tools available for conducting studies of landscape genetics.

Environmental risk factors for Lyme disease identified with geographic information systems.

GE Glass — 2006-08-02T14:56:10-00:00

Americal Journal of Public Health, Vol. 85, No. 7. (July 1995), pp. 944-948.

OBJECTIVES. A geographic information system was used to identify and locate residential environmental risk factors for Lyme disease. METHODS. Data were obtained for 53 environmental variables at the residences of Lyme disease case patients in Baltimore County from 1989 through 1990 and compared with data for randomly selected addresses. A risk model was generated combining the geographic information system with logistic regression analysis. The model was validated by comparing the distribution of cases in 1991 with another group of randomly selected addresses. RESULTS. In crude analyses, 11 environmental variables were associated with Lyme disease. In adjusted analyses, residence in forested areas (odds ratio [OR] = 3.7, 95% confidence interval [CI] = 1.2, 11.8), on specific soils (OR = 2.1, 95% CI = 1.0, 4.4), and in two regions of the county (OR = 3.5, 95% CI = 1.6, 7.4) (OR = 2.8, 95% CI = 1.0, 7.7) was associated with elevated risk of getting Lyme disease. Residence in highly developed regions was protective (OR = 0.3, 95% CI = 0.1, 1.0). The risk of Lyme disease in 1991 increased with risk categories defined from the 1989 through 1990 data. CONCLUSIONS. Combining a geographic information system with epidemiologic methods can be used to rapidly identify risk factors of zoonotic disease over large areas.