No abstract

Asthma is an inflammatory airways disease associated with intermittent respiratory symptoms, bronchial hyper-responsiveness (BHR) and reversible airflow obstruction and is phenotypically heterogeneous. Patterns of clustering and segregation analyses in asthma families have suggested a genetic component to asthma. Previous studies reported linkage of BHR and atopy to chromosomes 5q (refs 7-9), 6p (refs 10-12), 11q (refs 13-15), 14q (ref. 16), and 12q (ref. 17) using candidate gene approaches. However, the relative roles of these genes in the pathogenesis of asthma or atopy are difficult to assess outside of the context of a genome-wide search. One genome-wide search in atopic sib pairs has been reported, however, only 12% of their subjects had asthma. We conducted a genome-wide search in 140 families with > or = 2 asthmatic sibs, from three racial groups and report evidence for linkage to six novel regions: 5p15 (P = 0.0008) and 17p11.1-q11.2 (P = 0.0015) in African Americans; 11p15 (P = 0.0089) and 19q13 (P = 0.0013) in Caucasians; 2q33 (P = 0.0005) and 21q21 (P = 0.0040) in Hispanics. Evidence for linkage was also detected in five regions previously reported to be linked to asthma-associated phenotypes: 5q23-31 (P = 0.0187), 6p21.3-23 (P = 0.0129), 12q14-24.2 (P = 0.0042), 13q21.3-qter (P = 0.0014), and 14q11.2-13 (P = 0.0062) in Caucasians and 12q14-24.2 (P = 0.0260) in Hispanics.

No abstract

Due to genetic heterogeneity, phenocopies, incomplete penetrance, misdiagnosis, and unknown mode of inheritance, linkage studies of most complex diseases are unlikely to provide conclusive findings with unambiguously high lod scores. Typically, several marginally significant lod scores or elevated lod scores are observed in a genome-wide screen. However, it is usually difficult to differentiate these findings from false positives (type I errors). Two approaches are commonly used to guard against false positives: replication studies in independent samples and combined data analysis. In the current paper, we evaluated these two common approaches using simulated data where data from multiple groups were available and locations of disease genes were known. We found replication studies and combined data analysis performed similarly in terms of their ability to identify true and false positive linkages. Both approaches confirmed two true linkages and did not confirm any false positive linkages. The results also indicated that it is not appropriate to apply the criteria proposed for confirming significant evidence for linkage to confirm regions with only suggestive evidence for linkage. The current results support previous findings that parametric analysis using an incorrect genetic model can still identify a true linkage.

No abstract