The cutoff phenomenon in finite Markov chains.

Natural mixing processes modeled by Markov chains often show a sharp cutoff in their convergence to long-time behavior. This paper presents problems where the cutoff can be proved (card shuffling, the Ehrenfests' urn). It shows that chains with polynomial growth (drunkard's walk) do not show cutoffs. The best general understanding of such cutoffs (high multiplicity of second eigenvalues due to symmetry) is explored. Examples are given where the symmetry is broken but the cutoff phenomenon persists.

Stable and efficient gene transfer into the retina using an HIV-based lentiviral vector

The development of methods for efficient gene transfer to terminally differentiated retinal cells is important to study the function of the retina as well as for gene therapy of retinal diseases. We have developed a lentiviral vector system based on the HIV that can transduce terminally differentiated neurons of the brain in vivo. In this study, we have evaluated the ability of HIV vectors to transfer genes into retinal cells. An HIV vector containing a gene encoding the green fluorescent protein (GFP) was injected into the subretinal space of rat eyes. The GFP gene under the control of the cytomegalovirus promoter was efficiently expressed in both photoreceptor cells and retinal pigment epithelium. However, the use of the rhodopsin promoter resulted in expression predominantly in photoreceptor cells. Most successfully transduced eyes showed that photoreceptor cells in >80% of the area of whole retina expressed the GFP. The GFP expression persisted for at least 12 weeks with no apparent decrease. The efficient gene transfer into photoreceptor cells by HIV vectors will be useful for gene therapy of retinal diseases such as retinitis pigmentosa.

Integrin α6Aβ1 Induces CD81-dependent Cell Motility without Engaging the Extracellular Matrix Migration Substrate

It is well established that integrins and extracellular matrix (ECM) play key roles in cell migration, but the underlying mechanisms are poorly defined. We describe a novel mechanism whereby the integrin α6β1, a laminin receptor, can affect cell motility and induce migration onto ECM substrates with which it is not engaged. By using DNA-mediated gene transfer, we expressed the human integrin subunit α6A in murine embryonic stem (ES) cells. ES cells expressing α6A (ES6A) at the surface dimerized with endogenous β1, extended numerous filopodia and lamellipodia, and were intensely migratory in haptotactic assays on laminin (LN)-1. Transfected α6A was responsible for these effects, because cells transfected with control vector or α6B, a cytoplasmic domain α6 isoform, displayed compact morphology and no migration, like wild-type ES cells. The ES6A migratory phenotype persisted on fibronectin (Fn) and Ln-5. Adhesion inhibition assays indicated that α6β1 did not contribute detectably to adhesion to these substrates in ES cells. However, anti-α6 antibodies completely blocked migration of ES6A cells on Fn or Ln-5. Control experiments with monensin and anti-ECM antibodies indicated that this inhibition could not be explained by deposition of an α6β1 ligand (e.g., Ln-1) by ES cells. Cross-linking with secondary antibody overcame the inhibitory effect of anti-α6 antibodies, restoring migration or filopodia extension on Fn and Ln-5. Thus, to induce migration in ES cells, α6Aβ1 did not have to engage with an ECM ligand but likely participated in molecular interactions sensitive to anti-α6β1 antibody and mimicked by cross-linking. Antibodies to the tetraspanin CD81 inhibited α6Aβ1-induced migration but had no effect on ES cell adhesion. It is known that CD81 is physically associated with α6β1, therefore our results suggest a mechanism by which interactions between α6Aβ1 and CD81 may up-regulate cell motility, affecting migration mediated by other integrins.

Knowledge-based analysis of microarray gene expression data by using support vector machines

We introduce a method of functionally classifying genes by using gene expression data from DNA microarray hybridization experiments. The method is based on the theory of support vector machines (SVMs). SVMs are considered a supervised computer learning method because they exploit prior knowledge of gene function to identify unknown genes of similar function from expression data. SVMs avoid several problems associated with unsupervised clustering methods, such as hierarchical clustering and self-organizing maps. SVMs have many mathematical features that make them attractive for gene expression analysis, including their flexibility in choosing a similarity function, sparseness of solution when dealing with large data sets, the ability to handle large feature spaces, and the ability to identify outliers. We test several SVMs that use different similarity metrics, as well as some other supervised learning methods, and find that the SVMs best identify sets of genes with a common function using expression data. Finally, we use SVMs to predict functional roles for uncharacterized yeast ORFs based on their expression data.

Maximum likelihood estimation of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach

A maximum likelihood estimator based on the coalescent for unequal migration rates and different subpopulation sizes is developed. The method uses a Markov chain Monte Carlo approach to investigate possible genealogies with branch lengths and with migration events. Properties of the new method are shown by using simulated data from a four-population n-island model and a source–sink population model. Our estimation method as coded in migrate is tested against genetree; both programs deliver a very similar likelihood surface. The algorithm converges to the estimates fairly quickly, even when the Markov chain is started from unfavorable parameters. The method was used to estimate gene flow in the Nile valley by using mtDNA data from three human populations.

Extracellular Matrix Assembly in Diatoms (Bacillariophyceae)1 : III. Organization of Fucoglucuronogalactans within the Adhesive Stalks of Achnanthes longipes

Achnanthes longipes is a marine, biofouling diatom that adheres to surfaces via adhesive polymers extruded during motility or organized into structures called stalks that contain three distinct regions: the pad, shaft, and collar. Four monoclonal antibodies (AL.C1–AL.C4) and antibodies from two uncloned hybridomas (AL.E1 and AL.E2) were raised against the extracellular adhesives of A. longipes. Antibodies were screened against a hot-water-insoluble/hot-bicarbonate-soluble-fraction. The hot-water-insoluble/hot-bicarbonate-soluble fraction was fractionated to yield polymers in three size ranges: F1, ≥ 20,000,000 Mr; F2, ≅100,000 Mr; and F3, <10,000 Mr relative to dextran standards. The ≅100,000-Mr fraction consisted of highly sulfated (approximately 11%) fucoglucuronogalactans (FGGs) and low-sulfate (approximately 2%) FGGs, whereas F1 was composed of O-linked FGG (F2)-polypeptide (F3) complexes. AL.C1, AL.C2, AL.C4, AL.E1, and AL.E2 recognized carbohydrate complementary regions on FGGs, with antigenicity dependent on fucosyl-containing side chains. AL.C3 was unique in that it had a lower affinity for FGGs and did not label any portion of the shaft. Enzyme-linked immunosorbent assay and immunocytochemistry indicated that low-sulfate FGGs are expelled from pores surrounding the raphe terminus, creating the cylindrical outer layers of the shaft, and that highly sulfated FGGs are extruded from the raphe, forming the central core. Antibody-labeling patterns and other evidence indicated that the shaft central-core region is related to material exuded from the raphe during cell motility.

Efficient infection of a human T-cell line and of human primary peripheral blood leukocytes with a pseudotyped retrovirus vector.

Peripheral blood lymphocytes (PBLs) are an important target for gene transfer studies aimed at human gene therapy. However, no reproducibly efficient methods are currently available to transfer foreign, potentially therapeutic genes into these cells. While vectors derived from murine retroviruses have been the most widely used system, their low infection efficiency in lymphocytes has required prolonged in vitro culturing and selection after infection to obtain useful numbers of genetically modified cells. We previously reported that retroviral vectors pseudotyped with vesicular stomatitis G glycoprotein (VSV-G) envelope can infect a wide variety of cell types and can be concentrated to titers of greater than 10(9) infectious units/ml. In this present study, we examined the ability of amphotropic and pseudotyped vectors expressing a murine cell surface protein, B7-1, to infect the human T-cell line Jurkat or human blood lymphocytes. Limiting dilution analysis of transduced Jurkat cells demonstrated that the pseudotyped vector is significantly more efficient in infecting T cells than an amphotropic vector used at the same multiplicity of infection (moi). To identify the transduction efficiency on PBLs, we examined the levels of cell surface expression of the B7-1 surface marker 48 to 72 hr after infection. The transduction efficiency of PBLs with the pseudotyped vector increased linearly with increasing moi to a maximum of approximately 16-32% at an moi of 40. This relatively high efficiency of infection of a T-cell line and of blood lymphocytes with VSV-G pseudotyped virus demonstrates that such modified pseudotyped retrovirus vectors may be useful reagents for studies of gene therapy for a variety of genetic or neoplastic disorders.

Evidence for nonrandom hydrophobicity structures in protein chains.

The question of whether proteins originate from random sequences of amino acids is addressed. A statistical analysis is performed in terms of blocked and random walk values formed by binary hydrophobic assignments of the amino acids along the protein chains. Theoretical expectations of these variables from random distributions of hydrophobicities are compared with those obtained from functional proteins. The results, which are based upon proteins in the SWISS-PROT data base, convincingly show that the amino acid sequences in proteins differ from what is expected from random sequences in a statistically significant way. By performing Fourier transforms on the random walks, one obtains additional evidence for nonrandomness of the distributions. We have also analyzed results from a synthetic model containing only two amino acid types, hydrophobic and hydrophilic. With reasonable criteria on good folding properties in terms of thermodynamical and kinetic behavior, sequences that fold well are isolated. Performing the same statistical analysis on the sequences that fold well indicates similar deviations from randomness as for the functional proteins. The deviations from randomness can be interpreted as originating from anticorrelations in terms of an Ising spin model for the hydrophobicities. Our results, which differ from some previous investigations using other methods, might have impact on how permissive with respect to sequence specificity protein folding process is-only sequences with nonrandom hydrophobicity distributions fold well. Other distributions give rise to energy landscapes with poor folding properties and hence did not survive the evolution.

A technique for detecting matrix proteins in the crystalline spicule of the sea urchin embryo.

The presence of proteins associated with the CaCO3-containing biocrystals found in a wide variety of marine organisms is well established. In these organisms, including the primitive skeleton (spicule) of the sea urchin embryo, the structural and functional role of these proteins either in the biomineralization process or in control of the structural features of the biocrystals is unclear. Recently, one of the matrix proteins of the sea urchin spicule, SM 30, has been shown to contain a carbohydrate chain (the 1223 epitope) that has been implicated in the process whereby Ca2+ is deposited as CaCo3. Because an understanding of the localization of this protein, as well as other proteins found within the spicule, is central to understanding their function, we undertook to develop methods to localize spicule matrix proteins in intact spicules, using immunogold techniques and scanning electron microscopy. Gold particles indicative of this matrix glycoprotein could not be detected on the surface of spicules that had been isolated from embryo homogenates and treated with alkaline hypochlorite to remove any associated membranous material. However, when isolated spicules were etched for 2 min with dilute acetic acid (10 mM) to expose more internal regions of the crystal, SM 30 and perhaps other proteins bearing the 1223 carbohydrate epitope were detected in the calcite matrix. These results, indicating that these two antigens are widely distributed in the spicule, suggest that this technique should be applicable to any matrix protein for which antibodies are available.

Training and search methods for speech recognition.

Speech recognition involves three processes: extraction of acoustic indices from the speech signal, estimation of the probability that the observed index string was caused by a hypothesized utterance segment, and determination of the recognized utterance via a search among hypothesized alternatives. This paper is not concerned with the first process. Estimation of the probability of an index string involves a model of index production by any given utterance segment (e.g., a word). Hidden Markov models (HMMs) are used for this purpose [Makhoul, J. & Schwartz, R. (1995) Proc. Natl. Acad. Sci. USA 92, 9956-9963]. Their parameters are state transition probabilities and output probability distributions associated with the transitions. The Baum algorithm that obtains the values of these parameters from speech data via their successive reestimation will be described in this paper. The recognizer wishes to find the most probable utterance that could have caused the observed acoustic index string. That probability is the product of two factors: the probability that the utterance will produce the string and the probability that the speaker will wish to produce the utterance (the language model probability). Even if the vocabulary size is moderate, it is impossible to search for the utterance exhaustively. One practical algorithm is described [Viterbi, A. J. (1967) IEEE Trans. Inf. Theory IT-13, 260-267] that, given the index string, has a high likelihood of finding the most probable utterance.