Genomic outlier profile analysis: mixture models, null hypotheses, and nonparametric estimation

In most analyses of large-scale genomic data sets, differentialexpression analysis is typically assessed by testing for differencesin the mean of the distributions between 2 groups. A recentfinding by Tomlins and others (2005) is of a different typeof pattern of differential expression in which a fraction ofsamples in one group have overexpression relative to samplesin the other group. In this work, we describe a general mixturemodel framework for the assessment of this type of expression,called outlier profile analysis. We start by considering thesingle-gene situation and establishing results on identifiability.We propose 2 nonparametric estimation procedures that have naturallinks to familiar multiple testing procedures. We then developmultivariate extensions of this methodology to handle genome-widemeasurements. The proposed methodologies are compared usingsimulation studies as well as data from a prostate cancer geneexpression study.     

Balanced angular profile analysis

To evaluate current preferences and ethnic differences of female soft-tissue profiles, 71 profile photographs of famous female models were collected from Internet Web pages and divided into four groups (Korean, 22; Japanese, 15; Chinese, 16; and Western, 18). Eleven soft-tissue landmarks were recorded on each photograph and 16 angular measurements were made by using V-ceph (CyberMed, Inc., Seoul, Korea). Data from each group are presented to show the means, ranges, p and F values, standard deviations, and standard errors of each measurement. In addition, individual measurements for each group were compared with those of the other groups by one-way analysis of variance using a p value corrected for multivariable testing. Between-group mean value differences were calculated using a Tukey's studentized range test (HSD), at a significance level of p = 0.05. Most of the variables were similar in the groups. Significant between-group differences (p < 0.05) were found for angle of alar curvature point, profile convexity, interlabial contour, and nasolabial contour. In addition, we divided all data into two groups (Western and Asian). The t test (with significance level set to p = 0.05) was performed to compare the two. Significant between-group differences (p < 0.05) were found for angle of alar curvature, angle of labiale inferius, profile convexity, and lower lip projection angle, but no significant racial differences were found in terms of several profile angles. These findings suggest that point of ala curvature point, subnasale, and the labiale inferius of Asian models may differ from those of Western models. These peculiar angular patterns of Asian models led the authors to create a new characteristic angular concept, termed the "ethnic pyramid," which is composed of soft-tissue profile points of alar curvature point, subnasale, pronasale, and labiale inferius. This ethnic pyramid describes the characteristic patterns of the ethnic differences. The results of this study suggest that the soft-tissue profiles of famous female models have some common features but also show differences among ethnic groups and races. This simple method of profile analysis may provide aesthetic surgeons with a simple formula and reference data for creation and application of an attractive face. On the basis of their balanced angular profile analysis data, the authors suggest that appropriate and harmonious aesthetic operations reflecting these differences should be considered.     

Cancer outlier differential gene expression detection

We study statistical methods to detect cancer genes that are over- or down-expressed in some but not all samples in a disease group. This has proven useful in cancer studies where oncogenes are activated only in a small subset of samples. We propose the outlier robust t-statistic (ORT), which is intuitively motivated from the t-statistic, the most commonly used differential gene expression detection method. Using real and simulation studies, we compare the ORT to the recently proposed cancer outlier profile analysis (Tomlins and others, 2005) and the outlier sum statistic of Tibshirani and Hastie (2006). The proposed method often has more detection power and smaller false discovery rates. Supplementary information can be found at http://www.biostat.umn.edu/~baolin/research/ort.html.     

Microchip-based human serum atherogenic lipoprotein profile analysis

Owing to the mounting evidence of serum lipid changes in atherosclerosis, there has been increasing interest in developing new methods for analyzing atherogenic lipoprotein profiles. The separation of lipoprotein and lipoprotein subclasses has been demonstrated using a microchip capillary electrophoresis (CE) system [Chromatographia 74 (2011) 799–805]. In contrast to this previous study, the current report demonstrates that sdLDL peak efficiencies can be improved dramatically by adding gold nanoparticles (AuNPs) to the sample. Moreover, NBD C₆-ceramide was identified as a satisfactory dye for specific labeling and quantitation of individual serum lipoproteins. The accuracy of the method was evaluated by comparison with ultracentrifuge separated small, dense, low-density lipoprotein (sdLDL). A high correlation was observed between these two methods for sdLDL cholesterol. Lipid levels were investigated between atherosclerotic patients and healthy controls. The variation of serum atherogenic lipoprotein profiles for atherosclerotic patients pre- and post-treatment was assessed by microchip CE. This method has potential for the rapid and sensitive detection of different lipoprotein classes as well as their subclasses and, therefore, is suitable for routine clinical applications. Microchip-based atherogenic lipoprotein profile assays will greatly improve the analysis of risk factors in atherosclerosis and will provide useful information for monitoring the effect of therapies on atherosclerotic disease.     

Density estimation applications for outlier detection.

Nonparametric estimates of joint, conditional and marginal probability densities can be used to estimate the relative probability of a data point's recurrence. Outlier, unusual or abnormal values of a random variate tend to be those which are unlikely to recur. As part of an interactive graphical system, a procedure has been implemented which enables a biomedical researcher to view both the estimated probability and the numerical value of a data point's coordinates. This display circumvents the problem of interpreting a normal range in two or more dimensions and can thus be more easily generalized than most alternative outlier detection procedures.     

Shape based kinetic outlier detection in real-time PCR

Background  

Real-time PCR has recently become the technique of choice for absolute and relative nucleic acid quantification. The gold standard quantification method in real-time PCR assumes that the compared samples have similar PCR efficiency. However, many factors present in biological samples affect PCR kinetic, confounding quantification analysis. In this work we propose a new strategy to detect outlier samples, called SOD.     

The fine specificity of mannose-binding and galactose-binding lectins revealed using outlier motif analysis of glycan array data

Glycan-binding proteins are commonly used as analytical reagents to detect the levels of specific glycan structures in biological samples. A detailed knowledge of the specificities of glycan-binding proteins is required for properly interpreting their binding data. A powerful technology for characterizing glycan-binding specificity is the glycan array. However, the interpretation of glycan-array data can be difficult due to the complex fine specificities of certain glycan-binding proteins. We developed a systematic approach, called outlier-motif analysis, for extracting fine-specificity information from glycan-array data, and we applied the method to the study of four commonly used lectins: two mannose binders (concanavalin A and Lens culinaris) and two galactose binders (Bauhinia purpurea and peanut agglutinin). The study confirmed the known, primary specificity of each lectin and also revealed new insights into their binding preferences. Lens culinaris's main specificity may be non-terminal, α-linked mannose with a single linkage at its 2' carbon, which is more restricted than previous definitions. We found broader specificity for bauhinea purpurea (BPL) than previously reported, showing that BPL can bind terminal N-acetylgalactosamine (GalNAc) and penultimate β-linked galactose under certain limitations. Peanut agglutinin may bind terminal Galβ1,3Gal, a glycolipid motif, in addition to terminal Galβ1,3GalNAc, a common O-linked glycoprotein motif. These results could be used to more accurately interpret data obtained using these well-studied lectins. Furthermore, this study demonstrates a systematic and general approach for extracting fine-specificity information from glycan-array data.     

Time profile of hemin aggregation: an analysis

The status of hemin, whether monomeric or aggregated, appears to be a key factor in deciding its biological activity. The molecular basis for the tendency of hemin to aggregate and the factors that regulate it are not yet fully understood. We have investigated the time profile of aggregation of hemin in aqueous solutions and the effect of temperature on the process of aggregation. Aggregation increases with increase in temperature. The time profile data, as monitored by change in absorbance at 398 nm, fits a nonlinear equation with three time constants, suggesting a possibility of three processes. Interestingly, the variation of these three time constants with temperature are different.     

Genome analysis with the conditional multinomial distribution profile

The focus of the research is on the analysis of genome sequences. Based on the inter-nucleotide distance sequence, we propose the conditional multinomial distribution profile for the complete genomic sequence. These profiles can be used to define a very simple, computationally efficient, alignment-free, distance measure that reflects the evolutionary relationships between genomic sequences. We use this distance measure to classify chromosomes according to species of origin, to build the phylogenetic tree of 24 complete genome sequences of coronaviruses. Our results demonstrate the new method is powerful and efficient.     

Detecting outlier peptides in quantitative high-throughput mass spectrometry data

Quantitative high-throughput mass spectrometry has become an established tool to measure relative gene expression proteome-wide. The output of such an experiment usually consists of a list of expression ratios (fold changes) for several thousand proteins between two conditions. However, we observed that individual peptide fold changes may show a significantly different behavior than other peptides from the same protein and that these differences cannot be explained by imprecise measurements. Such outlier peptides can be the consequence of several technical (misidentifications, misquantifications) or biological (post-translational modifications, differential regulation of isoforms) reasons. We developed a method to detect outlier peptides in mass spectrometry data which is able to delineate imprecise measurements from real outlier peptides with high accuracy when the true difference is as small as 1.4 fold. We applied our method to experimental data and investigated the different technical and biological effects that result in outlier peptides. Our method will assist future research to reduce technical bias and can help to identify genes with differentially regulated protein isoforms in high throughput mass spectrometry data.     

Examining an outlier: molecular diversity in the cirripedia

Despite the typical assumption in studies of mitochondrial diversity that such data are useful for approximating population size and demography, studies of sequence diversity in mitochondrial DNA across the Metazoa have shown a surprising excess of rare alleles, a pattern associated either with strong selection or population growth. Previous work has shown that this bias toward an excess of rare alleles is typical across the Crustacea, and in particular, in the Cirripedia (barnacles). Here, we directly evaluate sequence data from studies of barnacle populations to ensure that inclusion of cryptic species is not the cause of this pattern. The results shown here reinforce previous studies that suggest caution in interpreting such patterns of allele frequencies, as they are likely to be influenced both by demographic changes and selection.