Michael Brown - Academia.edu (original) (raw)

Papers by Michael Brown

Ridden .Markov Models (HMMs) are applied t.0 the problems of statistical modeling, database searc... more Ridden .Markov Models (HMMs) are applied t.0 the problems of statistical modeling, database searching and multiple sequence alignment of protein families and protein domains. These methods are demonstrated on the globin family, the protein kinase catalytic domain, and the EF-hand calcium binding motif. In each case the parameters of an HMM are estimated from a training set of unaligned sequences. After the HMM is built, it is used to obtain a multiple alignment of all the training sequences. It is also used to search the . SWISS-PROT 22 database for other sequences. that are members of the given protein family, or contain the given domain. The Hi" produces multiple alignments of good quality that agree closely with the alignments produced by programs that incorporate threedimensional structural information. When employed in discrimination tests (by examining how closely the sequences in a database fit the globin, kinase and EF-hand HMMs), the '\ HMM is able to distinguish members of these families from non-members with a high degree ' of accuracy. Both the HMM and PROFILESEARCH ( a technique used to search for relationships between a protein sequence and multiply aligned sequences) perform better in these tests than PROSITE (a dictionary of sites and patterns in proteins). The HMM appecvs to have a slight advantage over PROFILESEARCH in terms of lower rates of false negatives and false positives, even though the HMM is trained using only unaligned sequences, whereas PROFILESEARCH requires aligned training sequences. Our results suggest the presence of an EF-hand calcium binding motif in a highly conserved and evolutionary preserved putative intracellular region of 155 residues in the a-l subunit of L-type calcium channels which play an important role in excitation-contraction coupling. This region has been suggested to contain the functional domains that are typical or essential for all L-type calcium channels regardless of whether they couple to ryanodine receptors, conduct ions or both. \ 8 Abbreviations HMM, hidden Marko,. ,,,dels; length from 130 to 170 residues (with few excep-' m, Expectetion-Maximization; ML. maximum tions) and two domains, the protein kinase catalytic like€ihood; MAP, maximum a posten'wi; hTL-score.

Proceedings of The National Academy of Sciences, 2000

We introduce a method of functionally classifying genes by using gene expression data from DNA mi... more 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.

Bioinformatics/computer Applications in The Biosciences, 1996

We present a method for condensing the information in multiple alignments of proteins into a mixt... more We present a method for condensing the information in multiple alignments of proteins into a mixture of Dirichlet densities over amino acid distributions. Dirichlet mixture densities are designed to be combined with observed amino acid frequencies to form estimates of expected amino acid probabilities at each position in a profile, hidden Markov model or other statistical model. These estimates give a statistical model greater generalization capacity, so that remotely related family members can be more reliably recognized by the model. This paper corrects the previously published formula for estimating these expected probabilities, and contains complete derivations of the Dirichlet mixture formulas, methods for optimizing the mixtures to match particular databases, and suggestions for efficient implementation.

A Bayesian method for estimating the amino acid distributions in the states of a hidden Markov mo... more A Bayesian method for estimating the amino acid distributions in the states of a hidden Markov model (HMM) for a protein family or the colunms of a multiple alignment of that family is introduced. This method uses Dirichlet mixture densities as priors over amino acid distributions. These mixture densities are determined from examination of previously constructed tlMMs or multiple alignments. It is shown that this Bayesian method can improve the quality of ItMMs produced from small training sets. Specific experiments on the EF-hand motif are reported, for which these priors are shown to produce HMMs with higher likelihood on unseen data, and fewer fal~ positives and false negatives in a database search task.

Journal of Clinical Investigation, 2002

Human Mutation, 1992

The low density lipoprotein (LDL) receptor is a cell surface transmembrane protein that mediates ... more The low density lipoprotein (LDL) receptor is a cell surface transmembrane protein that mediates the uptake and lysosomal degradation of plasma LDL, thereby providing cholesterol to cells. Mutations disrupting the function of this receptor produce autosomal dominant familial hypercholesterolemia (FH). Affected individuals have elevated plasma levels of LDL, which causes premature coronary atherosclerosis. To date, 71 mutations in the LDL receptor gene have been characterized at a molecular level. In this report, we describe 79 additional mutations and review the insights that all 150 mutations have provided into the structure/function relationship of the receptor protein and the clinical manifestations of FH. © 1992 Wiley-Liss, Inc.

Annual Review of Cell Biology, 1985

Cell, 1996

a transcription factor of the basic-helix-loop-helixleucine zipper family, a middle segment of ap... more a transcription factor of the basic-helix-loop-helixleucine zipper family, a middle segment of approxi-Joseph L. Goldstein, and Michael S. Brown mately 75 amino acids composed of two membrane-Department of Molecular Genetics spanning regions separated by a short hydrophilic loop University of Texas Southwestern Medical Center of 31 amino acids, and a carboxy-terminal segment of Dallas, Texas 75235 approximately 500 amino acids that plays a regulatory role . The SREBPs are oriented in the membranes of the endoplas-Summary mic reticulum (ER) and nuclear envelope in a hairpin fashion, so that the NH 2 -terminal and carboxy-terminal Through expression cloning we have isolated a cDNAsegments project into the cytosol and the 31 amino acid encoding SREBP cleavage-activating protein (SCAP), loop projects into the lumen (Hua et al., 1995). In cultured which regulates cholesterol metabolism by stimulatcells the two SREBPs act independently, and they are ing cleavage of transcription factors SREBP-1 and -2, apparently redundant. thereby releasing them from membranes. The cDNA

Molecular Cell, 2000

plasmic domain that projects into the ER lumen. When unfolded proteins accumulate in the ER, ATF6... more plasmic domain that projects into the ER lumen. When unfolded proteins accumulate in the ER, ATF6 is cleaved proteolytically to release the NH 2 -terminal domain, which enters the nucleus (Haze et al., 1999). There it activates transcription of at least three genes (GRP78, and calnexin) encoding chaperone proteins that Southwestern Medical Center restore the folding of proteins in the ER lumen. This Dallas, Texas 75390 unfolded protein response can be triggered by treatment † Department of Biological Sciences of cells with tunicamycin, which blocks an enzyme essen-Columbia University tial for the synthesis of asparagine-linked carbohydrates; New York, New York 10027 with thapsigargin, which depletes the ER of calcium; and with other agents that disrupt the folding of ER proteins (Pahl, 1999). The proteases that process ATF6, the sites Summary of cleavage within the protein, and the cellular locus of this cleavage have not yet been elucidated.

Cellular cholesterol homeostasis is controlled by sterol-regulated proteolysis of membrane-bound ... more Cellular cholesterol homeostasis is controlled by sterol-regulated proteolysis of membrane-bound transcription factors called sterol-regulatory element binding proteins (SREBPs). CPP32, a cysteine protease, was shown previously to cleave SREBP-1 and SREBP-2 in vitro at an aspartic acid between the basic helix-loop-helix leucine zipper domain and the first trans-membrane domain, liberating a transcriptionally active fragment. Here, we show that CPP32 exists in an inactive 32 kDa form in Chinese hamster ovary (CHO) cells. When apoptosis was induced with the protein kinase inhibitor staurosporine, CPP32 was cleaved to subunits of 20 and 10 kDa to form the active protease. Under these conditions membrane-bound SREBP-1 and SREBP-2 were both cleaved, and the transcriptionally active N-terminal fragments were found in nuclear extracts. Similar results were obtained in human U937 cells induced to undergo apoptosis by anti-Fas and etoposide. The apoptosis-induced cleavage of SREBPs was not suppressed by sterols, indicating that apoptosis-induced cleavage and sterol-regulated cleavage are mediated by different proteases. CHO cells expressing a mutant SREBP-2 with an Asp--> Ala mutation at the CPP32 cleavage site showed sterol-regulated cleavage but no apoptosis-induced cleavage. These data are consistent with the emerging concept that CPP32 is a central mediator in apoptosis. They also indicate that SREBPs, like poly (ADP) ribose polymerase, are cleaved by CPP32 during programmed cell death.

Cell, 1993

Sterol regulatory element I (SRE-1), a dscamer (5"ATC-ACCCCAC-3~ flanking the low density llpopro... more Sterol regulatory element I (SRE-1), a dscamer (5"ATC-ACCCCAC-3~ flanking the low density llpoprotein (LDL) receptor gene, activates transcription In sterol-deplated cells and is silenced by sterols. We report the cDNA cloning of human SREBP-1, a protein that binds SRE-1, activates transcription, and thereby mediates the final regulatory step in LDL metabolism. SREBP-1 contains a basic-helix-loop-helix-leucine zipper (bHLH-ZIP) motif, but It differs from other bHLH-ZIP proteins in its larger size (1147 amino acids) and target sequence. Instead of an inverted repeat (CANNTG), the target for all known bHLH-ZIP proteins, SRE-1 contains a direct repeat of CAC. Overexpression of SREBP-1 activates transcription of reporter genes containing SRE-1 in the absence (15-fold) and presence (90-fold) of sterols, abolishing sterol regulation. We suggest that SREBP-1 is regulated by an unknown factor that is overwhelmed when SREBP-1 is overexpressed. Understanding the regulation of SREBP-1 may be crucial for understanding the control of plasma cholesterol in humans.

Proceedings of The National Academy of Sciences, 1976

Cultured fibroblasts derived from patients with homozygous familial hypercholesterolemia, which l... more Cultured fibroblasts derived from patients with homozygous familial hypercholesterolemia, which lack functional low density lipoprotein.(IDL) receptors, fail to bind, take up, or degrade the lipoprotein with high affinity; therefore LDL-cholesterol is not made available for suppression of cholesterol synthesis or activation of cholesteryl ester formation. When LDL was given a positive charge by reaction with N,N-dimethyl4,3-propanediamine (cationized LDL), the rate of degradation of the lipoprotein was increased by more than 100 fold in the homozygous familial hypercholesterolemia fibroblasts. Degradation of cationized LDL was inhibited by chloroquine, suggesting that it occurred in cellular lysosomes. Although the cationized LDL entered the cell through a mechanism independent of the LDL receptor, the cholesterol liberated from the degradation of the lipoprotein became available for suppression of cholesterol synthesis and stimulation of cholesteryl ester formation in the homozygous familial hypercholesterolemia fibroblasts. The rate of degradation of albumin

Proceedings of The National Academy of Sciences, 2002

Human mtDNA shows striking regional variation, traditionally attributed to genetic drift. However... more Human mtDNA shows striking regional variation, traditionally attributed to genetic drift. However, it is not easy to account for the fact that only two mtDNA lineages (M and N) left Africa to colonize Eurasia and that lineages A, C, D, and G show a 5-fold enrichment from central Asia to Siberia. As an alternative to drift, natural selection might have enriched for certain mtDNA lineages as people migrated north into colder climates. To test this hypothesis we analyzed 104 complete mtDNA sequences from all global regions and lineages. African mtDNA variation did not significantly deviate from the standard neutral model, but European, Asian, and Siberian plus Native American variations did. Analysis of amino acid substitution mutations (nonsynonymous, Ka) versus neutral mutations (synonymous, Ks) (ka͞ks) for all 13 mtDNA proteincoding genes revealed that the ATP6 gene had the highest amino acid sequence variation of any human mtDNA gene, even though ATP6 is one of the more conserved mtDNA proteins. Comparison of the ka͞ks ratios for each mtDNA gene from the tropical, temperate, and arctic zones revealed that ATP6 was highly variable in the mtDNAs from the arctic zone, cytochrome b was particularly variable in the temperate zone, and cytochrome oxidase I was notably more variable in the tropics. Moreover, multiple amino acid changes found in ATP6, cytochrome b, and cytochrome oxidase I appeared to be functionally significant. From these analyses we conclude that selection may have played a role in shaping human regional mtDNA variation and that one of the selective influences was climate.