应用拟似性状方法在 COGA 数据集上研究染色体上Ⅰ型错误的关键值和变异。
Critical values and variation in type I error along chromosomes in the COGA dataset using the applied pseudo-trait method.
机构信息
Genometrics Section, Inherited Disease Research Branch, NHGRI/NIH, Baltimore, MD, USA.
出版信息
BMC Genet. 2005 Dec 30;6 Suppl 1(Suppl 1):S54. doi: 10.1186/1471-2156-6-S1-S54.
BACKGROUND
By analyzing a "pseudo-trait," a trait not linked or associated with any of the markers tested, the distribution of the test statistic under the null hypothesis can provide the critical value for the appropriate percentile of the distribution. In addition, the anecdotal observation that p-values tend to be more significant near the telomeres was investigated.
RESULTS
The applied pseudo-trait (APT) method was applied to the Affymetrix and Illumina SNPs in the Collaborative Study on the Genetics of Alcoholism dataset to determine appropriate critical values for regression of offspring on mid-parent (ROMP) and Haseman-Elston association and linkage analyses, investigating the occurrence of type I errors in different chromosomal locations, and the extent to which the critical values obtained depend on the type of pseudo-trait used.
CONCLUSION
On average, the 5 percentile critical values obtained for this study were less than the expected 0.05. The distribution of p-values does not seem to depend on chromosomal position for ROMP association analysis methods, but does in some cases for Haseman-Elston linkage analysis. Results vary with different pseudo-traits.
背景
通过分析一个“伪性状”，即与所测试的任何标记都没有关联或联系的性状，在零假设下，检验统计量的分布可以为分布的适当百分位数提供临界值。此外，还研究了一个轶事观察结果，即靠近端粒时 p 值往往更显著。
结果
将应用的伪性状（APT）方法应用于酒精中毒遗传学协作研究中的 Affymetrix 和 Illumina SNPs，以确定回归后代与中亲代（ROMP）和 Haseman-Elston 关联和连锁分析的适当临界值，调查不同染色体位置发生 I 型错误的情况，以及获得的临界值在多大程度上取决于所使用的伪性状类型。
结论
平均而言，本研究获得的 5%百分位临界值小于预期的 0.05。对于 ROMP 关联分析方法，p 值的分布似乎不取决于染色体位置，但在某些情况下，Haseman-Elston 连锁分析则不然。结果因不同的伪性状而异。
Zotero 插件