Search PubMed⌕ Search

Biomedical subjects

G D Stormo

Publications and source records attributed to G D Stormo.

At least 55 records · Page 3Linked to original sources
PubMed1994

Neural networks for determining protein specificity and multiple alignment of binding sites.

We use a quantitative definition of specificity to develop a neural network for the identification of common protein binding sites in a collection of unaligned DNA fragments. We demonstrate the equivalence of the method to maximizing Information Content of the aligned sites when simple models of the binding energy and the genome are employed. The network method subsumes those simple models and is capable of working with more complicated ones. This is demonstrated using a Markov model of the E. coli genome and a sampling method to approximate the partition function. A variation of Gibbs' sampling aids in avoiding local minima.

Bacterial Proteins↗
PubMed1994

Optimally parsing a sequence into different classes based on multiple types of evidence.

We consider the problem of parsing a sequence into different classes of subsequences. Two common examples are finding the exons and introns in genomic sequences and identifying the secondary structure domains of protein sequences. In each case there are various types of evidence that are relevant to the classification, but none are completely reliable, so we expect some weighted average of all the evidence to provide improved classifications. For example, in the problem of identifying coding regions in genomic DNA, the combined use of evidence such as codon bias and splice junction patterns can give more reliable predictions than either type of evidence alone. We show three main results: 1. For a given weighting of the evidence a dynamic programming algorithm returns the optimal parse and any number of sub-optimal parses. 2. For a given weighting of the evidence a dynamic programming algorithm determines the probability of the optimal parse and any number of sub-optimal parses under a natural Boltzmann-Gibbs distribution over the set of possible parses. 3. Given a set of sequences with known correct parses, a dynamic programming algorithm allows one to apply gradient descent to obtain the weights that maximize the probability of the correct parses of these sequences.

Algorithms↗
PubMed1993

Specificity of the Mnt protein. Independent effects of mutations at different positions in the operator.

The relative binding affinities of Mnt protein are determined for each possible base-pair at position 15 of the operator sequence, and for all combinations of G.C base-pairs at positions 15 and 17. The partitioning of each operator sequence is determined quantitatively with restriction enzymes. At position 15, the wild-type G.C base-pair provides the highest binding affinity but, unlike position 17, the primary distinction is between purine and pyrimidine bases on the top strand. The information content at position 15 is only about 0.16 bit. In comparison with previous measurements at position 17, it is determined that the interactions of the Mnt protein with positions 15 and 17 are independent, i.e. the specific binding energies for the two positions are additive. The relative binding affinities at position 17 are also determined in the background of a G to T mutation at position 5, the position equivalent to 17 on the other half of the symmetric operator. The relative affinities at position 17 are independent of whether position 5 is wild-type or mutant.

Base Sequence↗
PubMed1993

Identification of coding regions in genomic DNA sequences: an application of dynamic programming and neural networks.

Dynamic programming (DP) is applied to the problem of precisely identifying internal exons and introns in genomic DNA sequences. The program GeneParser first scores the sequence of interest for splice sites and for these intron- and exon-specific content measures: codon usage, local compositional complexity, 6-tuple frequency, length distribution and periodic asymmetry. This information is then organized for interpretation by DP. GeneParser employs the DP algorithm to enforce the constraints that introns and exons must be adjacent and non-overlapping and finds the highest scoring combination of introns and exons subject to these constraints. Weights for the various classification procedures are determined by training a simple feed-forward neural network to maximize the number of correct predictions. In a pilot study, the system has been trained on a set of 56 human gene fragments containing 150 internal exons in a total of 158,691 bps of genomic sequence. When tested against the training data, GeneParser precisely identifies 75% of the exons and correctly predicts 86% of coding nucleotides as coding while only 13% of non-exon bps were predicted to be coding. This corresponds to a correlation coefficient for exon prediction of 0.85. Because of the simplicity of the network weighting scheme, generalization performance is nearly as good as with the training set.

Algorithms↗
PubMed1993

Analysis of five presumptive protein-coding sequences clustered between the primosome genes, 41 and 61, of bacteriophages T4, T2, and T6.

In bacteriophage T4, there is a strong tendency for genes that encode interacting proteins to be clustered on the chromosome. There is 1.6 kb of DNA between the DNA helicase (gene 41) and the DNA primase (gene 61) genes of this virus. The DNA sequence of this region suggests that it contains five genes, designated as open reading frames (ORFs) 61.1 to 61.5, predicted to encode proteins ranging in size from 5.94 to 22.88 kDa. Are these ORFs actually genes? As one test, we compared the DNA sequence of this region in bacteriophages T2, T4, and T6 and found that ORFs 61.1, 61.3, 61.4, and 61.5 are highly conserved among the three closely related viruses. In contrast, ORF 61.2 is conserved between phages T4 and T6 yet is absent from phage T2, where it is replaced by another ORF, T2 ORF 61.2, which is not found in the T4 and T6 genomes. As a second, independent test for coding sequences, we calculated the codon base position preferences for all ORFs in this region that could encode proteins that contain at least 30 amino acids. Both the T4/T6 and T2 versions of ORF 61.2, as well as the other ORFs, have codon base position preferences that are indistinguishable from those of known T4 genes (coefficients of 0.81 to 0.94); the six other possible ORFs of at least 90 bp in this region are ruled out as genes by this test (coefficients less than zero). Thus, both evolutionary conservation and codon usage patterns lead us to conclude that ORFs 61.1 to 61.5 represent important protein-coding sequences for this family of bacteriophages. Because they are located between the genes that encode the two interacting proteins of the T4 primosome (DNA helicase plus DNA primase), one or more may function in DNA replication by modulating primosome function.

Amino Acid Sequence↗
PubMed1992

Identifying constraints on the higher-order structure of RNA: continued development and application of comparative sequence analysis methods.

Comparative sequence analysis addresses the problem of RNA folding and RNA structural diversity, and is responsible for determining the folding of many RNA molecules, including 5S, 16S, and 23S rRNAs, tRNA, RNAse P RNA, and Group I and II introns. Initially this method was utilized to fold these sequences into their secondary structures. More recently, this method has revealed numerous tertiary correlations, elucidating novel RNA structural motifs, several of which have been experimentally tested and verified, substantiating the general application of this approach. As successful as the comparative methods have been in elucidating higher-order structure, it is clear that additional structure constraints remain to be found. Deciphering such constraints requires more sensitive and rigorous protocols, in addition to RNA sequence datasets that contain additional phylogenetic diversity and an overall increase in the number of sequences. Various RNA databases, including the tRNA and rRNA sequence datasets, continue to grow in number as well as diversity. Described herein is the development of more rigorous comparative analysis protocols. Our initial development and applications on different RNA datasets have been very encouraging. Such analyses on tRNA, 16S and 23S rRNA are substantiating previously proposed associations and are now beginning to reveal additional constraints on these molecules. A subset of these involve several positions that correlate simultaneously with one another, implying units larger than a basepair can be under a phylogenetic constraint.

Base Sequence↗
PubMed1992

Splicing signals in Drosophila: intron size, information content, and consensus sequences.

A database of 209 Drosophila introns was extracted from Genbank (release number 64.0) and examined by a number of methods in order to characterize features that might serve as signals for messenger RNA splicing. A tight distribution of sizes was observed: while the smallest introns in the database are 51 nucleotides, more than half are less than 80 nucleotides in length, and most of these have lengths in the range of 59-67 nucleotides. Drosophila splice sites found in large and small introns differ in only minor ways from each other and from those found in vertebrate introns. However, larger introns have greater pyrimidine-richness in the region between 11 and 21 nucleotides upstream of 3' splice sites. The Drosophila branchpoint consensus matrix resembles C T A A T (in which branch formation occurs at the underlined A), and differs from the corresponding mammalian signal in the absence of G at the position immediately preceding the branchpoint. The distribution of occurrences of this sequence suggests a minimum distance between 5' splice sites and branchpoints of about 38 nucleotides, and a minimum distance between 3' splice sites and branchpoints of 15 nucleotides. The methods we have used detect no information in exon sequences other than in the few nucleotides immediately adjacent to the splice sites. However, Drosophila resembles many other species in that there is a discontinuity in A + T content between exons and introns, which are A + T rich.

Animals↗
PubMed1992

Expectation maximization algorithm for identifying protein-binding sites with variable lengths from unaligned DNA fragments.

An Expectation Maximization algorithm for identification of DNA binding sites is presented. The approach predicts the location of binding regions while allowing variable length spacers within the sites. In addition to predicting the most likely spacer length for a set of DNA fragments, the method identifies individual sites that differ in spacer size. No alignment of DNA sequences is necessary. The method is illustrated by application to 231 Escherichia coli DNA fragments known to contain promoters with variable spacings between their consensus regions. Maximum-likelihood tests of the differences between the spacing classes indicate that the consensus regions of the spacing classes are not distinct. Further tests suggest that several positions within the spacing region may contribute to promoter specificity.

Algorithms↗
PubMed1992

Translation initiation in Escherichia coli: sequences within the ribosome-binding site.

The translational roles of the Shine-Dalgarno sequence, the initiation codon, the space between them, and the second codon have been studied. The Shine-Dalgarno sequence UAAGGAGG initiated translation roughly four times more efficiently than did the shorter AAGGA sequence. Each Shine-Dalgarno sequence required a minimum distance to the initiation codon in order to drive translation; spacing, however, could be rather long. Initiation at AUG was more efficient than at GUG or UUG at each spacing examined; initiation at GUG was only slightly better than UUG. Translation was also affected by residues 3' to the initiation codon. The second codon can influence the rate of initiation, with the magnitude depending on the initiation codon. The data are consistent with a simple kinetic model in which a variety of rate constants contribute to the process of translation initiation.

Base Sequence↗
PubMed1991

Specificity of the Mnt protein determined by binding to randomized operators.

The relative binding affinities of Mnt protein from bacteriophage P22 are determined for each possible base pair at position 17 of the operator. These are determined from the partitioning of randomized operators into bound and unbound fractions; quantitation is provided by restriction enzyme analysis. Mnt protein is found to have an unusual specificity at this position: a C.G base pair (the wild-type operator) has the highest affinity, a G.C base pair has the lowest affinity, and both orientations of A.T base pairs are intermediate and nearly equivalent. A specific binding constant and specific binding free energy are defined and shown to be directly related to the information content of the operator sequences bound to the protein, taking into account the quantitative differences in binding affinities.

Bacteriophages↗
PubMed1991

Probing information content of DNA-binding sites.

An information content analysis of protein-binding sites gives a quantitative description of the specificity of the protein, independent of the mechanism of specificity. It gives useful information about the total specificity of the protein and about the individual positions within the binding sites. Information content is consistent with both thermodynamic and statistical analyses of specificity. When applied to a collection of known binding sites, the description provided may be limited by the sample size or by unknown constraints on those sites. Experimental procedures to determine the information content can give much more reliable measures. A large number of functional sites can be obtained from a much larger pool of randomized potential sites. Quantitative assays for the activity of different sites can be easily incorporated into the analysis, thereby increasing its sensitivity. Both in vitro and in vivo experiments are amenable to information content analysis.

Base Sequence↗
PubMed1991

Automated kinetic assay of beta-galactosidase activity.

An automated kinetic assay for beta-galactosidase activity in Escherichia coli was developed to permit the measurement of many independent samples simultaneously. Bacteria are grown, lysed from without (by adsorption of a high multiplicity of bacteriophage T4) and assayed in microtiter plates with 96 wells. Absorbance data are collected and analyzed by computer. The growth and lysis procedure, apparatus and software used in this assay can be used for other spectrophotometric enzyme assays.

Automation↗
PubMed1990

Sequence requirements of the hammerhead RNA self-cleavage reaction.

A previously well-characterized hammerhead catalytic RNA consisting of a 24-nucleotide substrate and a 19-nucleotide ribozyme was used to perform an extensive mutagenesis study. The cleavage rates of 21 different substrate mutations and 24 different ribozyme mutations were determined. Only one of the three phylogenetically conserved base pairs but all nine of the conserved single-stranded residues in the central core are needed for self cleavage. In most cases the mutations did not alter the ability of the hammerhead to assemble into a bimolecular complex. In the few cases where mutant hammerheads did not assemble, it appeared to be the result of the mutation stabilizing an alternate substrate or ribozyme secondary structure. All combinations of mutant substrate and mutant ribozyme were less active than the corresponding single mutations, suggesting that the hammerhead contains few, if any, replaceable tertiary interactions as are found in tRNA. The refined consensus hammerhead resulting from this work was used to identify potential hammerheads present in a variety of Escherichia coli gene sequences.

Base Sequence↗
PubMed1990

Consensus patterns in DNA.

Matrices can provide realistic representations of protein/DNA specificity. In many cases simple mononucleotide-based matrices are adequate representations, but more complex matrices may be needed for other cases. Unlike simple consensus sequences, matrices allow for different penalties to be assessed for different changes to a binding site, a property that is essential for accurate description of a binding site pattern. When only a collection of binding site sequences is known, the best representation for the pattern is an information content formulation, based on both thermodynamic and statistical considerations. Quantitative data on relative binding affinities may be used to determine matrices that provide a best fit to the data. Matrix representations also provide an efficient method of aligning multiple sequences to identify binding site patterns that they have in common.

Base Sequence↗
PubMed1990

Identification of consensus patterns in unaligned DNA sequences known to be functionally related.

We have developed a method for identifying consensus patterns in a set of unaligned DNA sequences known to bind a common protein or to have some other common biochemical function. The method is based on a matrix representation of binding site patterns. Each row of the matrix represents one of the four possible bases, each column represents one of the positions of the binding site and each element is determined by the frequency the indicated base occurs at the indicated position. The goal of the method is to find the most significant matrix--i.e. the one with the lowest probability of occurring by chance--out of all the matrices that can be formed from the set of related sequences. The reliability of the method improves with the number of sequences, while the time required increases only linearly with the number of sequences. To test this method, we analysed 11 DNA sequences containing promoters regulated by the Escherichia coli LexA protein. The matrices we found were consistent with the known consensus sequence, and could distinguish the generally accepted LexA binding sites from other DNA sequences.

Algorithms↗
PubMed1989

Excess information at bacteriophage T7 genomic promoters detected by a random cloning technique.

In our previous analysis of the information at binding sites on nucleic acids, we found that most of the sites examined contain the amount of information expected from their frequency in the genome. The sequences at bacteriophage T7 promoters are an exception, because they are far more conserved (35 bits of information content) than should be necessary to distinguish them from the background of the Escherichia coli genome (17 bits). To determine the information actually used by the T7 RNA polymerase, promoters were chemically synthesized with many variations and those that function well in an in vivo assay were sequenced. Our analysis shows that the polymerase uses 18 bits of information, so the sequences at phage genomic promoters have significantly more information than the polymerase needs. The excess may represent the binding site of another protein.

Base Composition↗
PubMed1989

Identifying protein-binding sites from unaligned DNA fragments.

The ability to determine important features within DNA sequences from the sequences alone is becoming essential as large-scale sequencing projects are being undertaken. We present a method that can be applied to the problem of identifying the recognition pattern for a DNA-binding protein given only a collection of sequenced DNA fragments, each known to contain somewhere within it a binding site for that protein. Information about the position or orientation of the binding sites within those fragments is not needed. The method compares the "information content" of a large number of possible binding site alignments to arrive at a matrix representation of the binding site pattern. The specificity of the protein is represented as a matrix, rather than a consensus sequence, allowing patterns that are typical of regulatory protein-binding sites to be identified. The reliability of the method improves as the number of sequences increases, but the time required increases only linearly with the number of sequences. An example, using known cAMP receptor protein-binding sites, illustrates the method.

Algorithms↗