Jiao Jin

Test Data Sets and Evaluation of Gene Prediction Programs on the Rice Genome

With several rice genome projects approaching completion gene prediction/finding by computer algorithms has become an urgent task. Two test sets were constructed by mapping the newly published 28,469 full-length KOME rice cDNA to the RGP BAC clone sequences of Oryza sativa ssp. japonica: a single-gene set of 550 sequences and a multi-gene set of 62 sequences with 271 genes. These data sets were used to evaluate five ab initio gene prediction programs: RiceHMM, GlimmerR, GeneMark, FGENSH and BGF. The predictions were compared on nucleotide, exon and whole gene structure levels using commonly accepted measures and several new measures. The test results show a progress in performance in chronological order. At the same time complementarity of the programs hints on the possibility of further improvement and on the feasibility of reaching better performance by combining several gene-finders.

Identification of climate factors related to human infection with avian influenza A H7N9 and H5N1 viruses in China.

Human influenza infections display a strongly seasonal pattern. However, whether H7N9 and H5N1 infections correlate with climate factors has not been examined. Here, we analyzed 350 cases of H7N9 infection and 47 cases of H5N1 infection. The spatial characteristics of these cases revealed that H5N1 infections mainly occurred in the South, Middle, and Northwest of China, while the occurrence of H7N9 was concentrated in coastal areas of East and South of China. Aside from spatial-temporal characteristics, the most adaptive meteorological conditions for the occurrence of human infections by these two viral subtypes were different. We found that H7N9 infections correlate with climate factors, especially temperature (TEM) and relative humidity (RHU), while H5N1 infections correlate with TEM and atmospheric pressure (PRS). Hence, we propose a risky window (TEM 4–14 °C and RHU 65–95%) for H7N9 infection and (TEM 2–22 °C and PRS 980-1025 kPa) for H5N1 infection. Our results represent the first step in determining the effects of climate factors on two different virus infections in China and provide warning guidelines for the future when provinces fall into the risky windows. These findings revealed integrated predictive meteorological factors rooted in statistic data that enable the establishment of preventive actions and precautionary measures against future outbreaks.

Robust discriminant analysis and its application to identify protein coding regions of rice genes.

Identification of protein coding regions is fundamentally a statistical pattern recognition problem. Discriminant analysis is a statistical technique for classifying a set of observations into predefined classes and it is useful to solve such problems. It is well known that outliers are present in virtually every data set in any application domain, and classical discriminant analysis methods (including linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA)) do not work well if the data set has outliers. In order to overcome the difficulty, the robust statistical method is used in this paper. We choose four different coding characters as discriminant variables and an approving result is presented by the method of robust discriminant analysis.

DISCRIMINANT ANALYSIS BASED ON STATISTICAL DEPTH

In the past two decades, many statistical depth functions seemed as powerful exploratory and inferential tools for multivariate data analysis have been presented. In this paper, a new depth function family that meets four properties mentioned in Zuo and Serfling (2000) is proposed. Then a classification rule based on the depth function family is proposed. The classification parameter b could be modified according to the type-I error α, and the estimator of b has the consistency and achieves the convergence rate n−1/2. With the help of the proper selection for depth family parameter c, the approach for discriminant analysis could minimize the type-II error β. A simulation study and a real data example compare the performance of the different discriminant methods.

Identification of risk factors in epidemiologic study based on ROC curve and network

This article proposes a new non-parametric approach for identification of risk factors and their correlations in epidemiologic study, in which investigation data may have high variations because of individual differences or correlated risk factors. First, based on classification information of high or low disease incidence, we estimate Receptor Operating Characteristic (ROC) curve of each risk factor. Then, through the difference between ROC curve of each factor and diagonal, we evaluate and screen for the important risk factors. In addition, based on the difference of ROC curves corresponding to any pair of factors, we define a new type of correlation matrix to measure their correlations with disease, and then use this matrix as adjacency matrix to construct a network as a visualization tool for exploring the structure among factors, which can be used to direct further studies. Finally, these methods are applied to analysis on water pollutants and gastrointestinal tumor, and analysis on gene expression data in tumor and normal colon tissue samples.

t-Type corrected-loss estimation for error-in-variable model

ABSTRACT In this article, we consider a linear model in which the covariates are measured with errors. We propose a t-type corrected-loss estimation of the covariate effect, when the measurement error follows the Laplace distribution. The proposed estimator is asymptotically normal. In practical studies, some outliers that diminish the robustness of the estimation occur. Simulation studies show that the estimators are resistant to vertical outliers and an application of 6-minute walk test is presented to show that the proposed method performs well.

Adaptive trimmed mean as a location estimate

The trimmed mean is one of the most common estimators of location for symmetrical distributions, whose effect depends on whether the trim rate matches the proportion of contaminated data. Based on the geometric characteristics of the curve of the trimmed variance function, the authors propose two kinds of adaptive trimmed mean algorithms. The accuracy of the estimators is compared with that of other often-used estimates, such as sample mean, trimmed mean, trimean, and median, by means of simulation method. The results show that the accuracy of the adaptive derivative optimization trimmed mean method is close to the optimum performance in case of medium contamination (the contamination rate is less than 50%). Under high contamination situation (the contamination rate equals 80%), the performance of the estimates is comparable to that of the median and is superior to other counterparts.