Variance-stabilizing transformations for two-color microarrays

MOTIVATION: Authors of several recent papers have independently introduced a family of transformations (the generalized-log family), which stabilizes the variance of microarray data up to the first order. However, for data from two-color arrays, tests for differential expression may require that the variance of the difference of transformed observations be constant, rather than that of the transformed observations themselves. RESULTS: We introduce a transformation within the generalized-log family which stabilizes, to the first order, the variance of the difference of transformed observations. We also introduce transformations from the 'started-log' and log-linear-hybrid families which provide good approximate variance stabilization of differences. Examples using control-control data show that any of these transformations may provide sufficient variance stabilization for practical applications, and all perform well compared to log ratios.     

THE LIFE CYCLE AND TOXICITY OF PFIESTERIA PISCICIDA REVISITED1

Despite use of excellent molecular techniques, Litaker et al. (2002) cannot provide insights about the life history of toxic Pfiesteria piscicida because they showed no data in support of having used toxic strains; rather they presented evidence that they used non‐inducible strains. Litaker et al. did not find amoeboid stages or a chrysophyte‐like cyst stage in several cultures and unequivocally concluded that the stages do not exist in all P. piscicida strains. Thus, they did not consider the tenet that absence of evidence does not constitute proof of absence. Apparent discrepancies between the research by Litaker et al. and previous research on Pfiesteria can be resolved as follows: First, Litaker et al. did not use toxic strains. We have reported findings (similar to Litaker et al.) showing few amoeboid transformations in non‐inducible strains, which manifest some but not all of the forms that have been documented in some toxic strains. We, and others, have documented active toxicity to fish, transformations to amoebae, and chrysophyte‐like cysts in some clonal toxic strains. Second, the data from several recent publications, which were available but not mentioned by Litaker et al. or by Coats (2002) in accompanying commentary, have verified P. piscicida amoebae, chrysophyte‐like cysts, and other stages in some toxic strains through a combination of approaches including PCR data from clonal cultures.     

Identification and distribution of protein families in 120 completed genomes using Gene3D

Using a new protocol, PFscape, we undertake a systematic identification of protein families and domain architectures in 120 complete genomes. PFscape clusters sequences into protein families using a Markov clustering algorithm (Enright et al., Nucleic Acids Res 2002;30:1575-1584) followed by complete linkage clustering according to sequence identity. Within each protein family, domains are recognized using a library of hidden Markov models comprising CATH structural and Pfam functional domains. Domain architectures are then determined using DomainFinder (Pearl et al., Protein Sci 2002;11:233-244) and the protein family and domain architecture data are amalgamated in the Gene3D database (Buchan et al., Genome Res 2002;12:503-514). Using Gene3D, we have investigated protein sequence space, the extent of structural annotation, and the distribution of different domain architectures in completed genomes from all kingdoms of life. As with earlier studies by other researchers, the distribution of domain families shows power-law behavior such that the largest 2,000 domain families can be mapped to approximately 70% of nonsingleton genome sequences; the remaining sequences are assigned to much smaller families. While approximately 50% of domain annotations within a genome are assigned to 219 universal domain families, a much smaller proportion (< 10%) of protein sequences are assigned to universal protein families. This supports the mosaic theory of evolution whereby domain duplication followed by domain shuffling gives rise to novel domain architectures that can expand the protein functional repertoire of an organism. Functional data (e.g. COG/KEGG/GO) integrated within Gene3D result in a comprehensive resource that is currently being used in structure genomics initiatives and can be accessed via http://www.biochem.ucl.ac.uk/bsm/cath/Gene3D/.     

Anomalous spreading speeds of cooperative recursion systems

This work presents an example of a cooperative system of truncated linear recursions in which the interaction between species causes one of the species to have an anomalous spreading speed. By this we mean that this species spreads at a speed which is strictly greater than its spreading speed in isolation from the other species and the speeds at which all the other species actually spread. An ecological implication of this example is discussed in Sect. 5. Our example shows that the formula for the fastest spreading speed given in Lemma 2.3 of our paper (Weinberger et al. in J Math Biol 45:183–218, 2002) is incorrect. However, we find an extra hypothesis under which the formula for the faster spreading speed given in (Weinberger et al. in J Math Biol 45:183–218, 2002) is valid. We also show that the hypotheses of all but one of the theorems of (Weinberger et al. in J Math Biol 45:183–218, 2002) whose proofs rely on Lemma 2.3 imply this extra hypothesis, so that all but one of the theorems of (Weinberger et al. in J Math Biol 45:183–218, 2002) and all the examples given there are valid as they stand.     

On the twin risk in autism

Autism is considered by many to be the most strongly genetically influenced multifactorial childhood psychiatric disorder. In the absence of any known gene or genes, the main support for this is derived from family and twin studies. Two recent studies (Greenberg et al. 2001; Betancur et al. 2002) suggested that the twinning process itself is an important risk factor in the development of autism. If true, this would have major consequences for the interpretation of twin studies. Both studies compared the number of affected twin pairs among affected sib pairs to expected values in two separate samples of multiplex families and reported a substantial and significant excess of twin pairs. Using data from our epidemiological study in Western Australia, we investigated the possibility of an increased rate of autism in twins. All children born between 1980 and 1995 with autism, Asperger syndrome, or pervasive developmental disorder not otherwise specified (PDD-NOS) were ascertained. Of the 465 children with a diagnosis, 14 were twin births (rate 30.0/1,000) compared to 9,640 children of multiple births out of a total of 386,637 births in Western Australia between 1980 and 1995 (twin rate weighted to number of children with autism or PDD per year 26.3/1,000). These data clearly do not support twinning as a substantial risk factor in the etiology of autism. We demonstrate that the high proportion of twins found in affected-sib-pair studies can be adequately explained by the high ratio of concordance rates in monozygotic (MZ) twins versus siblings and the distribution of family size in the population studied. Our results are in agreement with those of two similar studies by Croen et al. (2002) in California and Hultman et al. (2002) in Sweden.     

Linkage and association studies of QTL for nuclear families by mixed models

The transmission disequilibrium test (TDT) has been utilized to test the linkage and association between a genetic trait locus and a marker. Spielman et al. (1993) introduced TDT to test linkage between a qualitative trait and a marker in the presence of association. In the presence of linkage, TDT can be applied to test for association for fine mapping (Martin et al., 1997; Spielman and Ewens, 1996). In recent years, extensive research has been carried out on the TDT between a quantitative trait and a marker locus (Allison, 1997; Fan et al., 2002; George et al., 1999; Rabinowitz, 1997; Xiong et al., 1998; Zhu and Elston, 2000, 2001). The original TDT for both qualitative and quantitative traits requires unrelated offspring of heterozygous parents for analysis, and much research has been carried out to extend it to fit for different settings. For nuclear families with multiple offspring, one approach is to treat each child independently for analysis. Obviously, this may not be a valid method since offspring of one family are related to each other. Another approach is to select one offspring randomly from each family for analysis. However, with this method much information may be lost. Martin et al. (1997, 2000) constructed useful statistical tests to analyse the data for qualitative traits. In this paper, we propose to use mixed models to analyse sample data of nuclear families with multiple offspring for quantitative traits according to the models in Amos (1994). The method uses data of all offspring by taking into account their trait mean and variance-covariance structures, which contain all the effects of major gene locus, polygenic loci and environment. A test statistic based on mixed models is shown to be more powerful than the test statistic proposed by George et al. (1999) under moderate disequilibrium for nuclear families. Moreover, it has higher power than the TDT statistic which is constructed by randomly choosing a single offspring from each nuclear family.     

PS II model-based simulations of single turnover flash-induced transients of fluorescence yield monitored within the time domain of 100 ns–10 s on dark-adapted Chlorella pyrenoidosa cells

The set up described in Steffen et al. (Biochemistry 40:173-180, 2001) was used to monitor in the time domain from 100 ns to 10 s single turnover flash-induced transients of the normalized fluorescence yield (SFITFY) on dark-adapted cells of the thermophilic algae Chlorella pyrenoidosa Chick. Perfect data fit was achieved within the framework of a previously proposed model for the PS II reaction pattern (Lebedeva et al., Biophysics 47:968-980, 2002; Belyaeva et al., Biophysics 51:860-872, 2006) after its modification by taking into account nonradiative decay processes including nonphotochemical quenching due to time-dependent populations of P680(+*) and (3)Car. On the basis of data reported in the literature, a consistent set of rate constants was obtained for electron transfer at the donor and acceptor sides of PS II, pH in lumen and stroma, the initial redox state of plastoquinone pool and the rate of plastoquinone oxidation. The evaluation of the rate constant values of dissipative processes due to quenching by carotenoid triplets in antennae and P680(+*)Q(A)(-*) recombination as well as the initial state populations after excitation with a single laser flash are close to that outlined in (Steffen et al., Biochemistry 44:3123-3133, 2005a). The simulations based on the model of the PS II reaction pattern provide information on the time courses of population probabilities of different PS II states. We analyzed the maximum (F(m)(STF)) and minimum (F(0)) of the normalized FL yield dependence on the rate of the recombination processes (radiative and dissipative nonradiative) and of P680(+*) reduction. The developed PS II model provides a basis for theoretical comparative analyses of time-dependent fluorescence signals, observed at different photosynthetic samples under various conditions (e.g. presence of herbicides, other stress conditions, excitation with actinic pulses of different intensity, and duration).     

Comparative analysis of protein interaction networks

Recent advances in proteomics and computational biology have lead to a flood of protein interaction data and resulting interaction networks (e.g. (Gavin et al., 2002)). Here I first analyse the status and quality of parts lists (genes and proteins), then comparatively assess large-scale protein interaction data (von Mering et al., 2002) and finally try to identify biological meaningful units (e.g. pathways, cellular processes) within interaction networks that are derived from the conservation of gene neighborhood (Snel et al., 2002). Possible extensions of gene neighborhood analysis to eukaryotes (von Mering and Bork, 2002) will be discussed.     

《中国植物志》和《中国被子植物科属综论》所涉及"科"界定之比较

对采用A .Engler (1936 )被子植物分类系统的《中国植物志》和根据吴征镒等(2 0 0 2 )被子植物分类系统的《中国被子植物科属综论》中科的界定进行了比较。这不仅有助于掌握吴征镒等系统的观点 ,也有益于理解被子植物系统学中当前划分“科”的新趋向。     

Recognition of damaged DNA by Escherichia coli Fpg protein: insights from structural and kinetic data

Formamidopyrimidine-DNA glycosylase (Fpg) excises oxidized purines from damaged DNA. The recent determination of the three-dimensional structure of the covalent complex of DNA with Escherichia coli Fpg, obtained by reducing the Schiff base intermediate formed during the reaction [Gilboa et al., J. Biol. Chem. 277 (2002) 19811] has revealed a number of potential specific and non-specific interactions between Fpg and DNA. We analyze the structural data for Fpg in the light of the kinetic and thermodynamic data obtained by the method of stepwise increase in ligand complexity to estimate relative contributions of individual nucleotide units of lesion-containing DNA to its total affinity for this enzyme [Ishchenko et al., Biochemistry 41 (2002) 7540]. Stopped-flow kinetic analysis that has allowed the dissection of Fpg catalysis in time [Fedorova et al., Biochemistry 41 (2002) 1520] is also correlated with the structural data.     

Genetic signatures of strong recent positive selection at the lactase gene

In most human populations, the ability to digest lactose contained in milk usually disappears in childhood, but in European-derived populations, lactase activity frequently persists into adulthood (Scrimshaw and Murray 1988). It has been suggested (Cavalli-Sforza 1973; Hollox et al. 2001; Enattah et al. 2002; Poulter et al. 2003) that a selective advantage based on additional nutrition from dairy explains these genetically determined population differences (Simoons 1970; Kretchmer 1971; Scrimshaw and Murray 1988; Enattah et al. 2002), but formal population-genetics-based evidence of selection has not yet been provided. To assess the population-genetics evidence for selection, we typed 101 single-nucleotide polymorphisms covering 3.2 Mb around the lactase gene. In northern European-derived populations, two alleles that are tightly associated with lactase persistence (Enattah et al. 2002) uniquely mark a common (~77%) haplotype that extends largely undisrupted for >1 Mb. We provide two new lines of genetic evidence that this long, common haplotype arose rapidly due to recent selection: (1) by use of the traditional F(ST) measure and a novel test based on p(excess), we demonstrate large frequency differences among populations for the persistence-associated markers and for flanking markers throughout the haplotype, and (2) we show that the haplotype is unusually long, given its high frequency--a hallmark of recent selection. We estimate that strong selection occurred within the past 5,000-10,000 years, consistent with an advantage to lactase persistence in the setting of dairy farming; the signals of selection we observe are among the strongest yet seen for any gene in the genome.     

Direct interaction of Gas11 with microtubules: implications for the dynein regulatory complex

We previously described the Trypanin family of cytoskeleton-associated proteins that have been implicated in dynein regulation [Hill et al., J Biol Chem2000; 275(50):39369-39378; Hutchings et al., J Cell Biol2002;156(5):867-877; Rupp and Porter, J Cell Biol2003;162(1):47-57]. Trypanin from T. brucei is part of an evolutionarily conserved dynein regulatory system that is required for regulation of flagellar beat. In C. reinhardtii, the trypanin homologue (PF2) is part of an axonemal 'dynein regulatory complex' (DRC) that functions as a reversible inhibitor of axonemal dynein [Piperno et al., J Cell Biol1992;118(6):1455-1463; Gardner et al., J Cell Biol1994;127(5):1311-1325]. The DRC consists of an estimated seven polypeptides that are tightly associated with axonemal microtubules. Association with the axoneme is critical for DRC function, but the mechanism by which it attaches to the microtubule lattice is completely unknown. We demonstrate that Gas11, the mammalian trypanin/PF2 homologue, associates with microtubules in vitro and in vivo. Deletion analyses identified a novel microtubule-binding domain (GMAD) and a distinct region (IMAD) that attenuates Gas11-microtubule interactions. Using single-particle binding assays, we demonstrate that Gas11 directly binds microtubules and that the IMAD attenuates the interaction between GMAD and the microtubule. IMAD is able to function in either a cis- or trans-orientation with GMAD. The discovery that Gas11 provides a direct linkage to microtubules provides new mechanistic insight into the structural features of the dynein-regulatory complex.     

Semi-parametric differential expression analysis via partial mixture estimation

We develop an approach for microarray differential expression analysis, i.e. identifying genes whose expression levels differ between two or more groups. Current approaches to inference rely either on full parametric assumptions or on permutation-based techniques for sampling under the null distribution. In some situations, however, a full parametric model cannot be justified, or the sample size per group is too small for permutation methods to be valid. We propose a semi-parametric framework based on partial mixture estimation which only requires a parametric assumption for the null (equally expressed) distribution and can handle small sample sizes where permutation methods break down. We develop two novel improvements of Scott's minimum integrated square error criterion for partial mixture estimation [Scott, 2004a,b]. As a side benefit, we obtain interpretable and closed-form estimates for the proportion of EE genes. Pseudo-Bayesian and frequentist procedures for controlling the false discovery rate are given. Results from simulations and real datasets indicate that our approach can provide substantial advantages for small sample sizes over the SAM method of Tusher et al. [2001], the empirical Bayes procedure of Efron and Tibshirani [2002], the mixture of normals of Pan et al. [2003] and a t-test with p-value adjustment [Dudoit et al., 2003] to control the FDR [Benjamini and Hochberg, 1995].     

Modeling and E-M estimation of haplotype-specific relative risks from genotype data for a case-control study of unrelated individuals

The US National Cancer Institute has recently sponsored the formation of a Cohort Consortium (http://2002.cancer.gov/scpgenes.htm) to facilitate the pooling of data on very large numbers of people, concerning the effects of genes and environment on cancer incidence. One likely goal of these efforts will be generate a large population-based case-control series for which a number of candidate genes will be investigated using SNP haplotype as well as genotype analysis. The goal of this paper is to outline the issues involved in choosing a method of estimating haplotype-specific risk estimates for such data that is technically appropriate and yet attractive to epidemiologists who are already comfortable with odds ratios and logistic regression. Our interest is to develop and evaluate extensions of methods, based on haplotype imputation, that have been recently described (Schaid et al., Am J Hum Genet, 2002, and Zaykin et al., Hum Hered, 2002) as providing score tests of the null hypothesis of no effect of SNP haplotypes upon risk, which may be used for more complex tasks, such as providing confidence intervals, and tests of equivalence of haplotype-specific risks in two or more separate populations. In order to do so we (1) develop a cohort approach towards odds ratio analysis by expanding the E-M algorithm to provide maximum likelihood estimates of haplotype-specific odds ratios as well as genotype frequencies; (2) show how to correct the cohort approach, to give essentially unbiased estimates for population-based or nested case-control studies by incorporating the probability of selection as a case or control into the likelihood, based on a simplified model of case and control selection, and (3) finally, in an example data set (CYP17 and breast cancer, from the Multiethnic Cohort Study) we compare likelihood-based confidence interval estimates from the two methods with each other, and with the use of the single-imputation approach of Zaykin et al. applied under both null and alternative hypotheses. We conclude that so long as haplotypes are well predicted by SNP genotypes (we use the Rh2 criteria of Stram et al. [1]) the differences between the three methods are very small and in particular that the single imputation method may be expected to work extremely well.     

Nuclear import defect of human immunodeficiency virus type 1 DNA flap mutants is not dependent on the viral strain or target cell type

We have previously established, using human immunodeficiency virus type 1 (HIV-1) strain LAI, that the HIV-1 central DNA Flap acts as a cis determinant of viral genome nuclear import. Although the impact of the DNA Flap on nuclear import has already found numerous independent confirmations in the context of lentivirus vectors, it has been claimed that it may be nonessential for infectious virus strains LAI, YU-2 (J. D. Dvorin et al., J. Virol. 76:12087-12096, 2002), HXB2, and NL4-3 (A. Limon et al., J. Virol. 76:12078-12086, 2002). We conducted a detailed analysis of virus infectivity using the provirus clones provided by the authors and analogous target cells. In contrast to published data, our results show that all cPPT mutant viruses exhibit reduced infectivity corresponding to a nuclear import defect irrespective of the viral genetic background or target cell.     

Is synaptic homeostasis just wishful thinking?

The bone morphogenetic proteins (BMPs) are secreted polypeptides of the TGF-beta family, whose diverse functions include primary neural induction and the dorsoventral patterning of the neural tube. In this issue of Neuron, Aberle et al. (2002) and Marqués et al. present evidence that BMP receptors may also influence the development of synapses. The results suggest a novel mechanism for regulating neuronal growth and synaptic homeostasis during development.     

Functional connectivity metrics during stroke recovery

We explore functional connectivity in nine subjects measured with 1.5T fMRI-BOLD in a longitudinal study of recovery from unilateral stroke affecting the motor area (Small et al., 2002). We found that several measures of complexity of covariance matrices show strong correlations with behavioral measures of recovery. In Schmah et al. (2010), we applied Linear and Quadratic Discriminants (LD and QD) computed on a principal components (PC) subspace to classify the fMRI volumes into "early" and "late" sessions. We demonstrated excellent classification accuracy with QD but not LD, indicating that potentially important differences in functional connectivity exist between the early and late sessions. Motivated by Mclntosh et al. (2008), who showed that EEG brain-signal variability and behavioral performance both increased with age during development, we investigated complexity of the covariance matrix for this longitudinal stroke recovery data set. We used three complexity measures: the sphericity index described by Abdi (2010); "unsupervised dimensionality", which is the number of PCs that minimizes unsupervised generalization error of a covariance matrix (Hansen et al., 1999); and "QD dimensionality", which is the number of PCs that minimizes the classification accuracy of QD. Although these approaches measure different kinds of complexity, all showed strong correlations with one or more behavioral tests: nine-hole peg test, hand grip test and pinch test. We could not demonstrate that either sphericity or unsupervised dimensionality were significantly different for the "early" and "late" sessions using a paired Wilcoxon test. However, the amount of relative behavioral improvement was correlated with sphericity of the overall covariance matrix (pooled across all sessions), as well as with the divergence of the eigenspectra between the "early" and "late" covariance matrices. Complexity measures that use the number of PCs (which optimize QD classification or unsupervised generalization) were correlated with the behavioral performance of the final session, but not with the relative improvement. These are suggestive, but limited, results given the sample size, restricted behavioral measurements and older 1.5T BOLD data sets. Nevertheless, they indicate one potentially fruitful direction for future data-driven fMRI studies of stroke recovery in larger, better-characterized longitudinal stroke data sets recorded at higher field strength. Finally, we produced sensitivity maps (Kjems et al., 2002) corresponding to both linear and quadratic discriminants for the "early" vs. "late" classification. These maps measure the influence of each voxel on the class assignments for a given classifier. Differences between the scaled sensitivity maps for the linear and quadratic discriminants indicate brain regions involved in changes in functional connectivity. These regions are highly variable across subjects, but include the cerebellum and the motor area contralateral to the lesion.