Hao Mei
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Hao Mei has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Allison E. Ashley-Koch
Duke University Medical Center
1 shared publication - Deqiong Ma
Center for Human Genetics, Duke University Medical Center, Durham, NC 27710 USA
1 shared publication - Eden R. Martin
Center for Human Genetics, Duke University Medical Center, Durham, North Carolina, USA
1 shared publication - Elizabeth R. Hauser
VA Cooperative Studies Program Epidemiology Center—Durham, Department of Veterans Affairs, Durham, NC 27705, USA
1 shared publication - Michael A Schmidt
Center for Human Genetics, Duke University Medical Center, Durham NC, USA
1 shared publication - Silke Schmidt
Center for Human Genetics, Duke University Medical Center, Durham, NC, USA
1 shared publication
Latest Publications
- Open Access
- 30 December 2005
- Published by Springer Nature in BMC Genomic Data
- Vol. 6 (S1) , S148
Searching for epistatic interactions in nuclear families using conditional linkage analysis
Background: Genomic screens generally employ a single-locus strategy for linkage analysis, but this may have low power in the presence of epistasis. Ordered subsets analysis (OSA) is a method for conditional linkage analysis using continuous covariates. Methods: We used OSA to evaluate two-locus interactions in the simulated Genetic Analysis Workshop 14 dataset. We used all nuclear families ascertained by Aipotu, Karangar, and Danacaa. Using the single-nucleotide polymorphism map, multipoint affected-sibling-pair (ASP) linkage analysis was performed on all 100 replicates for each chromosome using SIBLINK. OSA was used to examine linkage on each chromosome using LOD scores at each 3-cM location on every other chromosome as covariates. Two methods were used to identify positive results: one searching across the entire covariate chromosome, the other conditioning on location of known disease loci. Results: Single-locus linkage analysis revealed very high LOD scores for disease loci D1 through D4, with mean LOD scores over 100 replicates ranging from 4.0 to 7.8. Although OSA did not obscure this linkage evidence, it did not detect the simulated interactions between any of the locus pairs. We found inflated type I error rates using the first OSA method, highlighting the need to correct for multiple comparisons. Therefore, using "null chromosome pairs" without simulated disease loci, we calculated a corrected alpha-level. Conclusion: We were unable to detect two-locus interactions using OSA. This may have been due to lack of incorporation of phenotypic subgroups, or because linkage evidence as summarized by LOD scores performs poorly as an OSA covariate. We found inflated type I error rates, but were able to calculate a corrected alpha-level for future analyses employing this strategy to search for two-locus interactions.
- 30 December 2005
- Published by Springer Nature in BMC Genomic Data
- Vol. 6 (S1) , S148
Open AccessSearching for epistatic interactions in nuclear families using conditional linkage analysis
Background: Genomic screens generally employ a single-locus strategy for linkage analysis, but this may have low power in the presence of epistasis. Ordered subsets analysis (OSA) is a method for conditional linkage analysis using continuous covariates. Methods: We used OSA to evaluate two-locus interactions in the simulated Genetic Analysis Workshop 14 dataset. We used all nuclear families ascertained by Aipotu, Karangar, and Danacaa. Using the single-nucleotide polymorphism map, multipoint affected-sibling-pair (ASP) linkage analysis was performed on all 100 replicates for each chromosome using SIBLINK. OSA was used to examine linkage on each chromosome using LOD scores at each 3-cM location on every other chromosome as covariates. Two methods were used to identify positive results: one searching across the entire covariate chromosome, the other conditioning on location of known disease loci. Results: Single-locus linkage analysis revealed very high LOD scores for disease loci D1 through D4, with mean LOD scores over 100 replicates ranging from 4.0 to 7.8. Although OSA did not obscure this linkage evidence, it did not detect the simulated interactions between any of the locus pairs. We found inflated type I error rates using the first OSA method, highlighting the need to correct for multiple comparisons. Therefore, using "null chromosome pairs" without simulated disease loci, we calculated a corrected alpha-level. Conclusion: We were unable to detect two-locus interactions using OSA. This may have been due to lack of incorporation of phenotypic subgroups, or because linkage evidence as summarized by LOD scores performs poorly as an OSA covariate. We found inflated type I error rates, but were able to calculate a corrected alpha-level for future analyses employing this strategy to search for two-locus interactions.
- Open Access
- 30 December 2005
- Published by Springer Nature in BMC Genomic Data
- Vol. 6 (S1) , S145
Extension of multifactor dimensionality reduction for identifying multilocus effects in the GAW14 simulated data
The multifactor dimensionality reduction (MDR) is a model-free approach that can identify gene × gene or gene × environment effects in a case-control study. Here we explore several modifications of the MDR method. We extended MDR to provide model selection without crossvalidation, and use a chi-square statistic as an alternative to prediction error (PE). We also modified the permutation test to provide different levels of stringency. The extended MDR (EMDR) includes three permutation tests (fixed, non-fixed, and omnibus) to obtain p-values of multilocus models. The goal of this study was to compare the different approaches implemented in the EMDR method and evaluate the ability to identify genetic effects in the Genetic Analysis Workshop 14 simulated data. We used three replicates from the simulated family data, generating matched pairs from family triads. The results showed: 1) chi-square and PE statistics give nearly consistent results; 2) results of EMDR without cross-validation matched that of EMDR with 10-fold cross-validation; 3) the fixed permutation test reports false-positive results in data from loci unrelated to the disease, but the non-fixed and omnibus permutation tests perform well in preventing false positives, with the omnibus test being the most conservative. We conclude that the non-cross-validation test can provide accurate results with the advantage of high efficiency compared to 10-cross-validation, and the non-fixed permutation test provides a good compromise between power and false-positive rate.
- 30 December 2005
- Published by Springer Nature in BMC Genomic Data
- Vol. 6 (S1) , S145
Open AccessExtension of multifactor dimensionality reduction for identifying multilocus effects in the GAW14 simulated data
The multifactor dimensionality reduction (MDR) is a model-free approach that can identify gene × gene or gene × environment effects in a case-control study. Here we explore several modifications of the MDR method. We extended MDR to provide model selection without crossvalidation, and use a chi-square statistic as an alternative to prediction error (PE). We also modified the permutation test to provide different levels of stringency. The extended MDR (EMDR) includes three permutation tests (fixed, non-fixed, and omnibus) to obtain p-values of multilocus models. The goal of this study was to compare the different approaches implemented in the EMDR method and evaluate the ability to identify genetic effects in the Genetic Analysis Workshop 14 simulated data. We used three replicates from the simulated family data, generating matched pairs from family triads. The results showed: 1) chi-square and PE statistics give nearly consistent results; 2) results of EMDR without cross-validation matched that of EMDR with 10-fold cross-validation; 3) the fixed permutation test reports false-positive results in data from loci unrelated to the disease, but the non-fixed and omnibus permutation tests perform well in preventing false positives, with the omnibus test being the most conservative. We conclude that the non-cross-validation test can provide accurate results with the advantage of high efficiency compared to 10-cross-validation, and the non-fixed permutation test provides a good compromise between power and false-positive rate.