Molecular classification of cancer types from microarray data using the combination of genetic algorithms and support vector machines

Simultaneous multiclass classification of tumor types is essential for future clinical implementations of microarray-based cancer diagnosis. In this study, we have combined genetic algorithms (GAs) and all paired support vector machines (SVMs) for multiclass cancer identification. The predictive features have been selected through iterative SVMs/GAs, and recursive feature elimination post-processing steps, leading to a very compact cancer-related predictive gene set. Leave-one-out cross-validations yielded accuracies of 87.93% for the eight-class and 85.19% for the fourteen-class cancer classifications, outperforming the results derived from previously published methods.     

Combining electron microscopic with x-ray crystallographic structures

Analgorithm has been developed for placing three-dimensional atomic structures into appropriately scaled cryoelectron microscopy maps. The first stage in this process is to conduct a three-dimensional angular search in which the center of gravity of an X-ray crystallographically determined structure is placed on a selected position in the cryoelectron microscopy map. The quality of the fit is measured by the sum of the density at each atomic position. The second stage is to refine the three angles and three translational parameters for the best (usually 25 to 100) fits. Useful criteria for this refinement include the sum of densities at atomic sites, the lack of atoms in negative or low density, the absence of atomic clashes between symmetry-related positions of the atomic structure, and the distances between identifiable features in the map and their positions on the fitted atomic structure. These refinements generally lead to a convergence of the originally chosen, top scoring fits to just a few (about 3 to 8) acceptable possibilities. Usually, the best remaining fit is clearly superior to any of the others.     

Some insights into protein structural class prediction

It has been quite clear that the success rate for predicting protein structural class can be improved significantly by using the algorithms that incorporate the coupling effect among different amino acid components of a protein. However, there is still a lot of confusion in understanding the relationship of these advanced algorithms, such as the least Mahalanobis distance algorithm, the component-coupled algorithm, and the Bayes decision rule. In this communication, a simple, rigorous derivation is provided to prove that the Bayes decision rule introduced recently for protein structural class prediction is completely the same as the earlier component-coupled algorithm. Meanwhile, it is also very clear from the derivative equations that the least Mahalanobis distance algorithm is an approximation of the component-coupled algorithm, also named as the covariant-discriminant algorithm introduced by Chou and Elrod in protein subcellular location prediction (Protein Engineering, 1999; 12:107-118). Clarification of the confusion will help use these powerful algorithms effectively and correctly interpret the results obtained by them, so as to conduce to the further development not only in the structural prediction area, but in some other relevant areas in protein science as well.     

中国濒危鱼类、两栖爬行类和哺乳类的地理分布格局与优先保护区域——基于《中国濒危动物红皮书》的分析

《中国濒危动物红皮书》依据我国动物所面临的濒危,对我国濒危动物的濒危等级划分,种群现状,致危因素和保护措施等进行了描述说明,首批收录了535种濒危动物。本文以《中国濒危动物红皮书》中所收录的352种濒危脊椎动物（不含鸟类（以下简记为“濒危植物”）为研究对象,整理统计了现有濒危物种的分布资料,在GIS支持下,对中国濒危动物的地理分布进行了研究。结果表明,中国濒危动物物种呈明显的集聚分布,最密集的地区是横断山区,海南岛,西双版纳和云贵高原;而在华北平原,内蒙古东部,黄土高原和东北平原等地区出现大片空白区。影响濒危动物分布的主要因素有热量和水分条件,地形条件等自然条件以及历史开发,人为破坏等人文条件。山地因地形屏障作用而保留了较多的古老物种,其较为复杂的环境因子也有益物种的生存,因而物种丰富度较高。大多数动物对于水分和热量的依赖性较强,因此水热条件对其分布的限制作用十分明显。人为破坏较为严重的地区,濒危物种稀少;污染,开荒等引起的环境问题对于现存动物的威胁很。运用Dobson排除算法得到云南勐腊县等9个县分布有168种濒危动物,占全国总数（海生种类及仅分布于台湾和香港的特有种除外）的51．5％,而其土地面积之和仅为全国陆地总面积的0．9％。而云南勐腊县等94个县市就分布有中国所有的濒危动物。这些地区是我国生物多样性保护应该优先考虑的地方。     

Analysis of failure time data with dependent interval censoring

This article develops a method for the analysis of screening data for which the chance of being screened is dependent on the event of interest (informative censoring). Because not all subjects make all screening visits, the data on the failure of interest is interval censored. We propose a model that will properly adjust for the dependence to obtain an unbiased estimate of the nonparametric failure time function, and we provide an extension for applying the method for estimation of the regression parameters from a (discrete time) proportional hazards regression model. The method is applied on a data set from an observational study of cytomegalovirus shedding in a population of HIV-infected subjects who participated in a trial conducted by the AIDS Clinical Trials Group.     

Efficient multipoint mapping: making use of dominant repulsion-phase markers

The paper is devoted to the problem of multipoint gene ordering with a particular focus on "dominance" complication that acts differently in conditions of coupling-phase and repulsion-phase markers. To solve the problem we split the dataset into two complementary subsets each containing shared codominant markers and dominant markers in the coupling-phase only. Multilocus ordering in the proposed algorithm is based on pairwise recombination frequencies and using the well-known travelling salesman problem (TSP) formalization. To obtain accurate results, we developed a multiphase algorithm that includes synchronized-marker ordering of two subsets assisted by re-sampling-based map verification, combining the resulting maps into an integrated map followed by verification of the integrated map. A new synchronized Evolution-Strategy discrete optimization algorithm was developed here for the proposed multilocus ordering approach in which common codominant markers facilitate stabilization of the marker order of the two complementary maps. High performance of the employed algorithm allows systematic treatment for the problem of verification of the obtained multilocus orders, based on computing-intensive bootstrap and jackknife technologies for detection and removing unreliable marker scores. The efficiency of the proposed algorithm was demonstrated on simulated and real data.Communicated by J.W. Snape     

Trophic determinants of hypopus induction in the stored-product mite Lepidoglyphus destructor (Acari: Astigmata)

A unique (synapomorphic) characteristic of astigmatic mites is the heteromorphic deuteronymph also called hypopus. It is a non-feeding and facultative instar between protonymph and tritonymph. The hypopus is adapted for dispersal and sometimes also for dormancy, as in Lepidoglyphus destructor. The experiments reveal a correlation between the composition of the foodstuff, the duration of development of homomorphic instars, the mortality of protonymphs and the production of hypopodes. As food quality decreases, development lasts longer, mortality increases and hypopodes are produced in greater numbers. Disadvantageous trophic conditions of varied chemical nature favour the induction of hypopodes. The experimental data show that hypopus incidences (as percentage individuals of a population) depend on the relative proportions of constituents of an ingested foodstuff. What matters is the ratio between nourishing foodstuff components and those that are of little or no nutritional value. When a certain ratio does not meet a presumed metabolically required level of nutrients a nutritional deficiency results and hypopus induction is triggered, provided that adequate genetic propensities for hypopus production are present (L. destructor is highly polymorphic for hypopus production). Specific key substances are apparently not involved, and composite properties of a foodstuff are crucial for hypopus induction. Decrease of food quality (not poor food per se) during the hypopus-inducible period (late larval to early protonymphal phase) promotes hypopus induction. The interpretation matches the ecological scene. When trophic deterioration of a patch habitat sets in, often as a result of overcrowding, conditions will eventually become untenable. As a response to incurring nutritional deficiencies the mites will induce hypopodes, which provide for escape from or survival at the decaying habitat patch. Experiments support the threshold model of quantitative genetics for hypopus expression as previously inferred from other experiments with L. destructor.     

Structure-based prediction of binding peptides to MHC class I molecules: application to a broad range of MHC alleles

Specific binding of antigenic peptides to major histocompatibility complex (MHC) class I molecules is a prerequisite for their recognition by cytotoxic T-cells. Prediction of MHC-binding peptides must therefore be incorporated in any predictive algorithm attempting to identify immunodominant T-cell epitopes, based on the amino acid sequence of the protein antigen. Development of predictive algorithms based on experimental binding data requires experimental testing of a very large number of peptides. A complementary approach relies on the structural conservation observed in crystallographically solved peptide-MHC complexes. By this approach, the peptide structure in the MHC groove is used as a template upon which peptide candidates are threaded, and their compatibility to bind is evaluated by statistical pairwise potentials. Our original algorithm based on this approach used the pairwise potential table of Miyazawa and Jernigan (Miyazawa S, Jernigan RL, 1996, J Mol Biol 256:623-644) and succeeded to correctly identify good binders only for MHC molecules with hydrophobic binding pockets, probably because of the high emphasis of hydrophobic interactions in this table. A recently developed pairwise potential table by Betancourt and Thirumalai (Betancourt MR, Thirumalai D, 1999, Protein Sci 8:361-369) that is based on the Miyazawa and Jernigan table describes the hydrophilic interactions more appropriately. In this paper, we demonstrate how the use of this table, together with a new definition of MHC contact residues by which only residues that contribute exclusively to sequence specific binding are included, allows the development of an improved algorithm that can be applied to a wide range of MHC class I alleles.     

Prediction of amino acid sequence from structure

We have developed a method for the prediction of an amino acid sequence that is compatible with a three-dimensional backbone structure. Using only a backbone structure of a protein as input, the algorithm is capable of designing sequences that closely resemble natural members of the protein family to which the template structure belongs. In general, the predicted sequences are shown to have multiple sequence profile scores that are dramatically higher than those of random sequences, and sometimes better than some of the natural sequences that make up the superfamily. As anticipated, highly conserved but poorly predicted residues are often those that contribute to the functional rather than structural properties of the protein. Overall, our analysis suggests that statistical profile scores of designed sequences are a novel and valuable figure of merit for assessing and improving protein design algorithms.     

Semiparametric regression analysis of interval-censored data

We propose a semiparametric approach to the proportional hazards regression analysis of interval-censored data. An EM algorithm based on an approximate likelihood leads to an M-step that involves maximizing a standard Cox partial likelihood to estimate regression coefficients and then using the Breslow estimator for the unknown baseline hazards. The E-step takes a particularly simple form because all incomplete data appear as linear terms in the complete-data log likelihood. The algorithm of Turnbull (1976, Journal of the Royal Statistical Society, Series B 38, 290-295) is used to determine times at which the hazard can take positive mass. We found multiple imputation to yield an easily computed variance estimate that appears to be more reliable than asymptotic methods with small to moderately sized data sets. In the right-censored survival setting, the approach reduces to the standard Cox proportional hazards analysis, while the algorithm reduces to the one suggested by Clayton and Cuzick (1985, Applied Statistics 34, 148-156). The method is illustrated on data from the breast cancer cosmetics trial, previously analyzed by Finkelstein (1986, Biometrics 42, 845-854) and several subsequent authors.