Krogh, A., Brown, M., Mian, I. S., Sj#olander, K., & Haussler, D. #1994#. Hidden Markov models in computational biology: Applications to protein modeling. JMB, 235, 1501#1531.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....sequences. In this case, each data instance in S is a sequence of variable length, with each position consisting of a residue a letter in a finite alphabet (A,C,G,T in the case of DNA sequences) To capture the characteristics of a class of sequences, we follow the approach of Haussler et al. [8] and represent the distribution over a set of multiplyaligned sequences using a profile hidden Markov model: For each concensus column of the multiple alignment, a match state models the distribution of the residues allowed in the column. An insert state and delete state at each column allow ....

.... H 1 )g =1 ; fPM i (O j H )g =1 ) where H and O are the state and residue variables, respectively, of the th column, and PM i represents the distribution induced by the profile M i , We note that a PAH with profile HMM CPMs resembles the mixture profile HMM approach of Krogh et al. [8]. Their approach consists of k different profile HMMs, each intended to capture a different subclass of proteins. However, there is no relation between the k profile HMMs, which can be arbitrarily different. By contrast, the PAH framework imposes an abstraction hierarchy over the different profile ....

[Article contains additional citation context not shown here]

A. Krogh, M. Brown, S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Mol. Biology, 235:1501--1531, 1994.

.... 1 Introduction One method of protein sequence analysis is the identi cation of homologous proteins proteins that share a common evolutionary history and have similar overall structure and function [5] Here we report on the use of SAM T98 [11, 16] a newly developed hidden Markov model (hmm) [13, 9] method for recognition of homologs with low sequence similarity, and how it fared in the foldrecognition section of the CASP3 experiment. Hmms combine the best aspects of weight matrices and local sequence alignment methods, and can be used to assign probabilities to proteins in database search ....

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. JMB, 235:1501{ 1531, Feb. 1994.

....piecewise dependent data. 1. Introduction PDD clustering has many applications in time signals including speaker recognition [1] 7] machine monitoring [8] clustering of EEG signals [9] and music clustering [10] Similar methods have also been applied in other areas, such as protein modeling [11]. PDD clustering must be used when there is a successive dependence between a group of data vectors and there no trained models of the data classes. In distance measure based clustering the goal is usually to minimize the overall distance. When a PDD is used longer segment supply more information ....

....segments leads to correct cluster despite the fact that the overall distance was very high. Knowledge or lack of knowledge about the boundaries of each data segment influences the problem s complexity. In several applications the segmentation is already given and only the labeling is missing [10] [11] or the segmentation can be found in advance [8] Consequently, in these cases all the available data for each segment should be employed. In other cases segment boundaries are unknown but might be estimated by combining a minimal duration constrain to ensure a sufficient statistics of each ....

A. Krogh, M. Brown, I. Saira Mian, K. Sjolander, and D. Haussler, "Hidden Markov models in computational biology applications to protein modeling," J. Mol. Biol., vol. 235, no. 5, pp. 1501-1531, 1994.

.... [5] protein secondary structure prediction [2] gene nding [11] recognition of transmembrane proteins [16] and characterization of biological sequence families by modeling how the biological sequences relate by substitutions, insertions, and deletions to the consensus sequence of the family [12]. Applications of hidden Markov models are often based on two fundamental questions. Given a hidden Markov model and a string we might want to determine the probability of the string under the model, i.e. determine the probability that the model has generated the string. This can be used for ....

A. Krogh, M. Brown, I. S. Mian, K. Sjlander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:15011531, 1994.

....order to find a good match to a sequence with unknown tertiary structure. Similar sequences tend to have similar structure, and in fact there are broad categories of structure that most proteins (or portions thereof) fall into. In an e#ort to exploit these structural categories, Krogh 4 et al. [37] build probabilistic models for amino acid sequences conditional on structural classes. These models are built up from known structure sequence pairs, and then used to infer a likely structural class for a novel amino acid sequence. Thus, for example, a stochastic model is built for the sequence ....

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: applications to protein modeling. Technical Report, Computer and Information Sciences, University of California, Santa Cruz, 1992.

....to stripe data across multiple server disks reaching 97 of the combined write capacity of multiple nodes. 1. Introduction Data intensive applications constitute an increasing share of high performance computing (HPC) An increasing number of applications in domains such as genomics proteomics [1,2,3,4], astrophysics [5] geophysics [6] computational neuroscience [7] or volume rendering [8] need to archive, retrieve, and process increasingly large datasets. These applications are prime candidates for Grid computing [9] as they involve remote access and extensive computation to many data ....

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler, Hidden Markov models in computational biology: Applications to protein modeling, JMB 235:1501-1531, 1994.

....no function or structure is known yet. They serve as a basis for learning models which are used to detect sub sequences (called domains or motifs) which are known to be related to a particular biochemical function. Such models range from complex probabilistic models based on hidden Markov models [11, 5] to purely syntactic models, like regular expressions, describing characteristic sub sequences [1] However, since the databases are constantly increasing and updated, the learning procedure of these models must be easy and of low complexity. 3.1 Protein domains detection with variable Markov ....

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501-1531, 1994.

....can use [10, 31] as starting points for a broader overview of the subject. EM like algorithms for HMM s were introduced in [4, 3, 2, 24] The EM family of algorithms was introduced in great generality in [9] work on HMM s also appears in the econometrics [13, 23] as well as in the biological [22] literature. These references are merely starting points; the literature is very extensive. As already mentioned, the EM segmentation algorithm used here is a variation of algorithms which are well established in the field of speech recognition; for example see [18, 19] Taking into account the ....

A. Krogh et al. "Hidden Markov models in computational biology: applications to protein model- ing". J. Mol. Biol., vol.235, pp.150-1531, 1994.

....In principle, we can use any probabilistic model for the CPM as long as it defines a probability distribution over the state space. We have recently [14] applied this approach to the substantially more complex problem of clustering proteins based on their amino acid sequence using profile HMMs [11]. Acknowledgements. We thank Nir Friedman for useful comments. This work was supported by NSF Grant ACI 0082554 under the NSF ITR program, and by the Sloan Foundation. Eran Segal was also supported by a Stanford Graduate Fellowship (SGF) ....

A. Krogh, M. Brown, S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Mol. Biology, 235:1501--1531, 1994.

....a set of generative models, each of which is trained over one class of data. An unknown sequence is evaluated by each generative model to determine which class (or protein family) is most likely to have generated the protein sequence. These models include approaches such as hidden Markov models [25, 10, 2, 24], probabilistic suf x trees [3, 1] and pro les [13] In contrast, discriminative models are trained over data containing multiple classes and directly classify an unknown sequence into a class. Discriminative models have been successfully applied to protein homology using support vector machines ....

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:15011531, 1994.

....is used as a distance metric to form an initial partitioning of the time series for the subsequent HMM based clustering procedure. Finite mixtures of HMMs have also been studied by a number of researchers. Similar to mixtures of Markov chains, the EM algorithm can also be used for HMM mixtures [17, 18, 19, 20]. To trade accuracy for e#ciency, the k means algorithm (used in [21] and the rival penalized competitive learning (RPCL) algorithm (used in [22] have also been used in place of EM. The number of clusters can be determined using Monte Carlo cross validation [19] or information criteria such as ....

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: applications to protein modeling. Journal of Molecular Biology, 235(5):1501--1531, 1994.

....to the best of our knowledge, none of them test performance with the leave half out criterion. Various machine learning techniques have been applied to the protein structure prediction problem. The two main approaches are neural nets (e.g. 47, 67, 59] and hidden Markov models (e.g. [53, 9]) Both of these approaches require adequate data on the target motif, since there is a training session on sequences that are known to contain the target motif. Our approach differs from these methods since it does not require well analyzed data on the Test examples Initial parameters ....

A. Krogh, M. Brown, S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Technical Report UCSC-CRL93 -32, University of California at Santa Cruz, 1993.

....rather than the biological meaning of the sequences, that is, we have to deal with strings having the A, C, G, U symbols. These strings present structured regions caused by palindrome pairs and non structured regions where any long term relation can be found. Stochastic local models such as HMMs [3, 2], n grams, etc, have been used in order to model RNA sequences. However, these models are not the most appropriate because palindrome paired positions are treated independently due to the restriction of the structure of these local models. Thus, a more powerful class of languages is needed to ....

A.Krogh, M.Brown, I.S.Mian, K.Sjolander, D.Haussler. Hidden Markov Models in Computational Biology: Applications to Protein Modeling, Journal Molecular Biology, 235:15011531, February 1994.

....[59] proposes a heuristic simultaneously aligning the sequences and building a phylogeny for the sequences; Bucka Lassen et al. 23] propose a method for combining many multiple alignments into an improved alignment. One of the most successful and popular heuristics, introduced by Krogh et al. [80], is using profile hidden Markov models to generate an alignment. 2.2.2 Hidden Markov Models A Markov chain is a sequence of symbols or states q i 0 q i 1 . q i j . where the probability of observing a specific state in the j th position only depends on the state observed in the j 1 st ....

....Markov models is for annotating a sequence s; given a path through a hidden Markov model that generates s we can annotate the characters of s by the states outputting them. In the following we will make this concept clear by developing the architecture known as profile hidden Markov models, cf. [80]. Assume we have some sequences that all descend from a common ancestral sequence AG of two characters, cf. figure 2.5. If we start out by only modelling positions corresponding to characters present in the ancestral sequence, we can A phylogeny for a set of sequences is a tree showing ....

[Article contains additional citation context not shown here]

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

....between the sequences into account. Hein in [82] presents a heuristic which simultaneously attempts to infer and use the evolutionary relationships between members of a sequence family to guide the construction of a multiple alignment of the members of the sequence family. Krogh et al. in [111] present a popular and successful heuristic for computing multiple alignments, which use profile hidden Markov models to describe the relationships between members of a sequence family. We return to profile hidden Markov models in Section 2.2.2. A multiple alignment of a set of strings is useful ....

....in the late 1980 s and early 1990 s. Since then they have found many applications, e.g. modeling of DNA [38] protein secondary structure prediction [15] gene prediction [110] and recognition of transmembrane proteins [175] Probably the most popular application, introduced by Krogh et al. in [111], is to use profile hidden Markov models to characterize a sequence family by modeling how the sequences relate by substitutions, insertions and deletions to the consensus sequence of the family. The prefix profile is because profile hidden Markov models address the same problem as profiles of ....

[Article contains additional citation context not shown here]

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

....1989 by Churchill [5] Durbin et al. 6] and Eddy [7, 8] are good overviews of the use of hidden Markov models in computational biology. One of the most popular applications is to use them to characterise sequence families by using so called pro le hidden Markov models introduced by Krogh et al. [15]. For a pro le hidden Markov model the probability of a given sequence indicates how likely it is that the sequence is a member of the modelled sequence family, and the most likely path for a given sequence corresponds to an alignment of the sequence against the modelled sequence family. An ....

....scheme that captures the characteristics of a sequence family, in the sense that the score peaks around members of the family. Pro les are useful when searching for unknown members of a sequence family and several methods have been used to construct and use pro les [10, 16, 21] Krogh et al. [15] realized that simple hidden Markov models, which they called pro le hidden Markov models, were able to capture all other pro le methods. The states of a pro le hidden Markov model are divided into match , insert and delete states. Figure 1 illustrates the transition structure of a simple ....

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501-1531, 1994.

....and Bioethics. Bioinformatics at UCSC UC Santa Cruz is well known in the field of bioinformatics for pioneering work on applications of hidden Markov models (HMMs) and stochastic context free grammars to biological sequence data and for the development of software tools for sequence analysis [4], 5] 6] 7] Our efforts include participation in the international Human Genome Project [8] development of the gene finding software chosen to annotate the human and Drosophila (fly) genomes [9] success in international protein structure prediction contests [10] 11] and creation of a ....

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler, "Hidden Markov models in computational biology: Applications to protein modeling," JMB, vol. 235, pp. 1501--1531, Feb. 1994.

....Foundation probability re ects a sequence likely to belong to the class of sequences being modeled, instead of formalisms only distinguishing sequences as either belonging to the class being modeled or not. The two most widely used grammatical models in bioinformatics are hidden Markov models [1, 3, 8, 9, 16] and stochastic context free grammars [7, 14, 15] though other models have also been proposed [13, 18] These two types of stochastic models were originally developed as tools for speech recognition (see [2, 12] One can identify hidden Markov models as a stochastic version of regular languages ....

A. Krogh, M. Brown, I. S. Mian, K. Sjlander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:15011531, 1994.

....models where a model is trained over one class of data. An unknown protein sequence is evaluated by each protein family model to determine which protein family is most likely to have generated the protein sequence. These models include approaches such as hidden Markov modelbased approaches [24, 9, 2, 23], probabilistic suffix tree based approaches [3, 1] and profiles based approaches [12] Recently discriminative models have been applied to protein homology and approaches have included using support vector machines [21] and sparse Markov transducer approach [10] In discriminative models, the ....

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

.... Research Center, www.birc.dk, funded by the University of Aarhus Research Foundation applications, e.g. modeling of DNA sequences [5] protein secondary structure prediction [2] gene nding [11] recognition of transmembrane proteins [16] and characterization of biological sequence families [12]. Applications of HMMs are often based on two fundamental questions. Given an HMM and a string we might want to determine the probability of the string under the model, i.e. the probability that the model has generated the string. This can be used for classi cation of the string as either ....

A. Krogh, M. Brown, I. S. Mian, K. Sjlander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Jour. Mol. Biol, 235:15011531, 1994.

No context found.

Krogh, A.; Brown, M.; Mian, I. S.; Sj#olander, K.; and Haussler, D. 1994. Hidden Markov models in computational biology: Applications to protein modeling. JMB 235:1501#1531.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. JMB, 235:1501-1531, February 1994.

No context found.

Krogh, A., Brown, M., Mian, I. S., Sj#olander, K., & Haussler, D. #1994#. Hidden Markov models in computational biology: Applications to protein modeling. JMB, 235, 1501#1531.

No context found.

Krogh, A., Brown, M., Mian, I. S., Sjolander, K., & Haussler, D. (1994). Hidden Markov models in computational biology: Applications to protein modeling. J. Mol. Biol. 235, 1501-1531.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler, Hidden Markov models in computational biology: Applications to protein modeling, Journal of Molecular Biology, 235:1501--1531, February 1994.

No context found.

A. Krogh, B. Brown, I. S. Mian, K. Sjolander,, D. Haussler. Hidden Markov models in computational biology: applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. J. Mol. Biol. 235: 1501-1531, 1994.

No context found.

Krogh, A. (1994). Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994. CLOSING REMARKS xxxv

No context found.

A. Krogh, Hidden Markov Models in computational biology: Applications to protein modelling. Journal of Molecular Biology, 235:1501--1531.

No context found.

Krogh, A., Brown, M., Mian, I., Sjolander, K., and Haussler, D., Hidden markov models in computational biology: applications to protein modeling, J. Mol. Biol., 235:1501--1531, 1994.

No context found.

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501-- 1531, 1994.

No context found.

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler, "Hidden Markov Models in Computational Biology Applications to Protein Modeling," Journal of Molecular Biology, vol. 235, no. 5, pp. 1501--1531, Feb. 1994.

No context found.

Krogh, A., Brown, M., Mian, I. S., Sjolander, K., and Haussler, D. (1993). Hidden Markov models in computational biology: Applications to protein modeling. Technical Report UCSC-CRL-93-32, University of California, Santa Cruz.

No context found.

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501--1531, 1994.

No context found.

Krogh, A. et al. (1994). Hidden Markov Models in Computational Biology: Applications to Protein Modelling, Journal of Molecular Biology 235, 1501-1531.

No context found.

Krogh A, Brown M, Mian IS, Sjolander K, Haussler D. Hidden Markov models in computational biology: Applications to protein modeling. J Mol Biol 1994;235:1501--1531.

No context found.

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: applications to protein modeling. JMB, 235:1501-1531, 1994.

No context found.

Krogh, A., Brown, M., Mian, I. S., Sjolander, K., and Haussler, D. (1993). Hidden Markov models in computational biology: Applications to protein modeling. Technical Report UCSC-CRL-93-32, University of California, Santa Cruz.

No context found.

Krogh,A., Brown,M., Mian,I. S., Sjolander,K., and Haussler,D. (1994) Hidden markov models in computational biology: Applications to protein modeling. JMB, 235:1501--1531.

No context found.

Krogh, A., Brown, M., Mian, I. S., Sjo lander, K. & Haussler, D. (1996). Hidden Markov models in computational biology: application to protein modeling. J. Mol. Biol. ###, 1501-1531.

No context found.

A. Krogh, M. Brown, I.S. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: applications to protein modeling. J Mol Biol., 235:1501--1531, 1994.

No context found.

Anders Krogh, Michael Brown, I. Saira Mian, Kimmen Sjolander, and David Haussler. Hidden Markov models in computational biology: Applications to protein modeling. Technical Report UCSC-CRL-93-32, University of California at Santa Cruz, 1993.

No context found.

A. Krogh, M. Brown, I. Mian, K. Sjolander, and D. Haussler. Hidden markov models in computational biology: Applications to protein modeling. Journal of Molecular Biology, 235:1501-- 1531, 1994.

No context found.

A. Krogh, M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. Hidden Markov models in computational biology: Applications to protein modeling. J. Mol. Biol., 235:1501--1531, Feb 1994.

No context found.

A. Krogh, M. Brown, I. Saira Mian, K. Sjolander, and D. Haussler, "Hidden Markov models in computational biology applications to protein modeling," J. Mol. Biol., vol. 235, no. 5, pp. 1501-1531, February 1994.

No context found.

Krogh, A., M. Brown, I. S. Mian, K. Sjolander, and D. Haussler. 1994. Hidden Markov models in computational biology: applications to protein modeling. J. Mol. Biol. 235:1501.

No context found.

Krogh,A., Brown,M., Mian,I.S. Sjlander,K. and Haussler,D. (1994) Hidden Markov models in computational biology: application to protein modeling. J. Mol. Biol. 235, 1501--1531.

No context found.

2. Krogh, A., M. Brown, I. S. Mian, K. Sjolander & D. Haussler (1994). "Hidden Markov Models in Computational Biology: Applications to Protein Modeling." J. Mol. Biol. 235: 1501-1531. 13

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC