Dataset Information

Evaluating statistical methods using plasmode data sets in the age of massive public databases: an illustration using false discovery rates.

ABSTRACT: Plasmode is a term coined several years ago to describe data sets that are derived from real data but for which some truth is known. Omic techniques, most especially microarray and genomewide association studies, have catalyzed a new zeitgeist of data sharing that is making data and data sets publicly available on an unprecedented scale. Coupling such data resources with a science of plasmode use would allow statistical methodologists to vet proposed techniques empirically (as opposed to only theoretically) and with data that are by definition realistic and representative. We illustrate the technique of empirical statistics by consideration of a common task when analyzing high dimensional data: the simultaneous testing of hundreds or thousands of hypotheses to determine which, if any, show statistical significance warranting follow-on research. The now-common practice of multiple testing in high dimensional experiment (HDE) settings has generated new methods for detecting statistically significant results. Although such methods have heretofore been subject to comparative performance analysis using simulated data, simulating data that realistically reflect data from an actual HDE remains a challenge. We describe a simulation procedure using actual data from an HDE where some truth regarding parameters of interest is known. We use the procedure to compare estimates for the proportion of true null hypotheses, the false discovery rate (FDR), and a local version of FDR obtained from 15 different statistical methods.

SUBMITTER: Gadbury GL

PROVIDER: S-EPMC2409977 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Genome-wide detection of quantitative trait loci (QTL) hotspots underlying variation in many molecular and phenotypic traits has been a key step in various biological studies since the QTL hotspots are highly informative and can be linked to the genes for the quantitative traits. Several statistical methods have been proposed to detect QTL hotspots. These hotspot detection methods rely heavily on permutation tests performed on summarized QTL data or individual-level data (with genotypes and phenotypes) from the genetical genomics experiments. In this article, we propose a statistical procedure for QTL hotspot detection by using the summarized QTL (interval) data collected in public web-accessible databases. First, a simple statistical method based on the uniform distribution is derived to convert the QTL interval data into the expected QTL frequency (EQF) matrix. And then, to account for the correlation structure among traits, the QTL for correlated traits are grouped together into the same categories to form a reduced EQF matrix. Furthermore, a permutation algorithm on the EQF elements or on the QTL intervals is developed to compute a sliding scale of EQF thresholds, ranging from strict to liberal, for assessing the significance of QTL hotspots. With grouping, much stricter thresholds can be obtained to avoid the detection of spurious hotspots. Real example analysis and simulation study are carried out to illustrate our procedure, evaluate the performances and compare with other methods. It shows that our procedure can control the genome-wide error rates at the target levels, provide appropriate thresholds for correlated data and is comparable to the methods using individual-level data in hotspot detection. Depending on the thresholds used, more than 100 hotspots are detected in GRAMENE rice database. We also perform a genome-wide comparative analysis of the detected hotspots and the known genes collected in the Rice Q-TARO database. The comparative analysis reveals that the hotspots and genes are conformable in the sense that they co-localize closely and are functionally related to relevant traits. Our statistical procedure can provide a framework for exploring the networks among QTL hotspots, genes and quantitative traits in biological studies. The R codes that produce both numerical and graphical outputs of QTL hotspot detection in the genome are available on the worldwide web http://www.stat.sinica.edu.tw/chkao/.

Dataset Information

Evaluating statistical methods using plasmode data sets in the age of massive public databases: an illustration using false discovery rates.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets