Speech perception at the interface of neurobiology and linguistics

Speech perception consists of a set of computations that take continuously varying acoustic waveforms as input and generate discrete representations that make contact with the lexical representations stored in long-term memory as output. Because the perceptual objects that are recognized by the speech perception enter into subsequent linguistic computation, the format that is used for lexical representation and processing fundamentally constrains the speech perceptual processes. Consequently, theories of speech perception must, at some level, be tightly linked to theories of lexical representation. Minimally, speech perception must yield representations that smoothly and rapidly interface with stored lexical items. Adopting the perspective of Marr, we argue and provide neurobiological and psychophysical evidence for the following research programme. First, at the implementational level, speech perception is a multi-time resolution process, with perceptual analyses occurring concurrently on at least two time scales (approx. 20-80 ms, approx. 150-300 ms), commensurate with (sub)segmental and syllabic analyses, respectively. Second, at the algorithmic level, we suggest that perception proceeds on the basis of internal forward models, or uses an 'analysis-by-synthesis' approach. Third, at the computational level (in the sense of Marr), the theory of lexical representation that we adopt is principally informed by phonological research and assumes that words are represented in the mental lexicon in terms of sequences of discrete segments composed of distinctive features. One important goal of the research programme is to develop linking hypotheses between putative neurobiological primitives (e.g. temporal primitives) and those primitives derived from linguistic inquiry, to arrive ultimately at a biologically sensible and theoretically satisfying model of representation and computation in speech.     

Bayesian hierarchically weighted finite mixture models for samples of distributions

Finite mixtures of Gaussian distributions are known to provide an accurate approximation to any unknown density. Motivated by DNA repair studies in which data are collected for samples of cells from different individuals, we propose a class of hierarchically weighted finite mixture models. The modeling framework incorporates a collection of k Gaussian basis distributions, with the individual-specific response densities expressed as mixtures of these bases. To allow heterogeneity among individuals and predictor effects, we model the mixture weights, while treating the basis distributions as unknown but common to all distributions. This results in a flexible hierarchical model for samples of distributions. We consider analysis of variance-type structures and a parsimonious latent factor representation, which leads to simplified inferences on non-Gaussian covariance structures. Methods for posterior computation are developed, and the model is used to select genetic predictors of baseline DNA damage, susceptibility to induced damage, and rate of repair.     

Chaos game representation for comparison of whole genomes

Background  

Chaos game representation of genome sequences has been used for visual representation of genome sequence patterns as well as alignment-free comparisons of sequences based on oligonucleotide frequencies. However the potential of this representation for making alignment-based comparisons of whole genome sequences has not been exploited.     

Weighted relative entropy for alignment-free sequence comparison based on Markov model

In this paper, we introduce a probabilistic measure for computing the similarity between two biological sequences without alignment. The computation of the similarity measure is based on the Kullback-Leibler divergence of two constructed Markov models. We firstly validate the method on clustering nine chromosomes from three species. Secondly, we give the result of similarity search based on our new method. We lastly apply the measure to the construction of phylogenetic tree of 48 HEV genome sequences. Our results indicate that the weighted relative entropy is an efficient and powerful alignment-free measure for the analysis of sequences in the genomic scale.     

A simple iterative approach to parameter optimization.

Various bioinformatics problems require optimizing several different properties simultaneously. For example, in the protein threading problem, a scoring function combines the values for different parameters of possible sequence-to-structure alignments into a single score to allow for unambiguous optimization. In this context, an essential question is how each property should be weighted. As the native structures are known for some sequences, a partial ordering on optimal alignments to other structures, e.g., derived from structural comparisons, may be used to adjust the weights. To resolve the arising interdependence of weights and computed solutions, we propose a heuristic approach: iterating the computation of solutions (here, threading alignments) given the weights and the estimation of optimal weights of the scoring function given these solutions via systematic calibration methods. For our application (i.e., threading), this iterative approach results in structurally meaningful weights that significantly improve performance on both the training and the test data sets. In addition, the optimized parameters show significant improvements on the recognition rate for a grossly enlarged comprehensive benchmark, a modified recognition protocol as well as modified alignment types (local instead of global and profiles instead of single sequences). These results show the general validity of the optimized weights for the given threading program and the associated scoring contributions.     

On adaptation: A reduction of the Kauffman-Levin model to a problem in graph theory and its consequences

It is shown that complex adaptations are best modelled as discrete processes represented on directed weighted graphs. Such a representation captures the idea that problems of adaptation in evolutionary biology are problems in a discrete space, something that the conventional representations using continuous adaptive landscapes does not. Further, this representation allows the utilization of well-known algorithms for the computation of several biologically interesting results such as the accessibility of one allele from another by a specified number of point mutations, the accessibility of alleles at a local maximum of fitness, the accessibility of the allele with the globally maximum fitness, etc. A reduction of a model due to Kauffman and Levin to such a representation is explicitly carried out and it is shown how this reduction clarifies the biological questions that are of interest.Thanks are due to William Wimsatt, James F. Crow, and the referees for Biology and Philosophy for comments on an earlier version of this paper. Remarks by members of the audience, especially Abner Shimony, of a seminar at Boston University, February 19, 1988, were also very helpful. The diagrams were prepared with the assistance of Tracy Lubas.     

A simple method to analyze the similarity of biological sequences based on the fuzzy theory

In this paper, we propose a simple method to analyze the similarity of biological sequences. By taking the average contents of biological sequences and their information entropies as the variables, the fuzzy method is used to cluster them. From the results of application, it finds that the method is relatively easy and rapid. Unlike other methods such as the graphical representation methods, which is usually very complex to compute some invariants of matric derived from graphical representation, our method pays more attention to the information of biological sequences themselves. Especially with the help of the software (SPSS), it seems to be very convenient. Therefore, it may be used to study the new biological sequences such as their evolution relationship and structures.     

A weighted measure for the similarity analysis of DNA sequences

Here we propose a weighted measure for the similarity analysis of DNA sequences. It is based on LZ complexity and (0,1) characteristic sequences of DNA sequences. This weighted measure enables biologists to extract similarity information from biological sequences according to their requirements. For example, by this weighted measure, one can obtain either the full similarity information or a similarity analysis from a given biological aspect. Moreover, the length of DNA sequence is not problematic. The application of the weighted measure to the similarity analysis of β-globin genes from nine species shows its flexibility.     

Super pairwise alignment (SPA): an efficient approach to global alignment for homologous sequences.

Sequence analysis is the basis of bioinformatics, while sequence alignment is a fundamental task for sequence analysis. The widely used alignment algorithm, Dynamic Programming, though generating optimal alignment, takes too much time due to its high computation complexity O(N(2)). In order to reduce computation complexity without sacrificing too much accuracy, we have developed a new approach to align two homologous sequences. The new approach presented here, adopting our novel algorithm which combines the methods of probabilistic and combinatorial analysis, reduces the computation complexity to as low as O(N). The computation speed by our program is at least 15 times faster than traditional pairwise alignment algorithms without a loss of much accuracy. We hence named the algorithm Super Pairwise Alignment (SPA). The pairwise alignment execution program based on SPA and the detailed results of the aligned sequences discussed in this article are available upon request.     

Continuous Distributed Representation of Biological Sequences for Deep Proteomics and Genomics

We introduce a new representation and feature extraction method for biological sequences. Named bio-vectors (BioVec) to refer to biological sequences in general with protein-vectors (ProtVec) for proteins (amino-acid sequences) and gene-vectors (GeneVec) for gene sequences, this representation can be widely used in applications of deep learning in proteomics and genomics. In the present paper, we focus on protein-vectors that can be utilized in a wide array of bioinformatics investigations such as family classification, protein visualization, structure prediction, disordered protein identification, and protein-protein interaction prediction. In this method, we adopt artificial neural network approaches and represent a protein sequence with a single dense n-dimensional vector. To evaluate this method, we apply it in classification of 324,018 protein sequences obtained from Swiss-Prot belonging to 7,027 protein families, where an average family classification accuracy of 93%±0.06% is obtained, outperforming existing family classification methods. In addition, we use ProtVec representation to predict disordered proteins from structured proteins. Two databases of disordered sequences are used: the DisProt database as well as a database featuring the disordered regions of nucleoporins rich with phenylalanine-glycine repeats (FG-Nups). Using support vector machine classifiers, FG-Nup sequences are distinguished from structured protein sequences found in Protein Data Bank (PDB) with a 99.8% accuracy, and unstructured DisProt sequences are differentiated from structured DisProt sequences with 100.0% accuracy. These results indicate that by only providing sequence data for various proteins into this model, accurate information about protein structure can be determined. Importantly, this model needs to be trained only once and can then be applied to extract a comprehensive set of information regarding proteins of interest. Moreover, this representation can be considered as pre-training for various applications of deep learning in bioinformatics. The related data is available at Life Language Processing Website: http://llp.berkeley.edu and Harvard Dataverse: http://dx.doi.org/10.7910/DVN/JMFHTN.     

Detection of secondary structure elements in proteins by hydrophobic cluster analysis.

Hydrophobic cluster analysis (HCA) is a protein sequence comparison method based on alpha-helical representations of the sequences where the size, shape and orientation of the clusters of hydrophobic residues are primarily compared. The effectiveness of HCA has been suggested to originate from its potential ability to focus on the residues forming the hydrophobic core of globular proteins. We have addressed the robustness of the bidimensional representation used for HCA in its ability to detect the regular secondary structure elements of proteins. Various parameters have been studied such as those governing cluster size and limits, the hydrophobic residues constituting the clusters as well as the potential shift of the cluster positions with respect to the position of the regular secondary structure elements. The following results have been found to support the alpha-helical bidimensional representation used in HCA: (i) there is a positive correlation (clearly above background noise) between the hydrophobic clusters and the regular secondary structure elements in proteins; (ii) the hydrophobic clusters are centred on the regular secondary structure elements; (iii) the pitch of the helical representation which gives the best correspondence is that of an alpha-helix. The correspondence between hydrophobic clusters and regular secondary structure elements suggests a way to implement variable gap penalties during the automatic alignment of protein sequences.     

Exploring Approximate Bayesian Computation for inferring recent demographic history with genomic markers in nonmodel species

Approximate Bayesian computation (ABC) is widely used to infer demographic history of populations and species using DNA markers. Genomic markers can now be developed for nonmodel species using reduced representation library (RRL) sequencing methods that select a fraction of the genome using targeted sequence capture or restriction enzymes (genotyping‐by‐sequencing, GBS). We explored the influence of marker number and length, knowledge of gametic phase, and tradeoffs between sample size and sequencing depth on the quality of demographic inferences performed with ABC. We focused on two‐population models of recent spatial expansion with varying numbers of unknown parameters. Performing ABC on simulated data sets with known parameter values, we found that the timing of a recent spatial expansion event could be precisely estimated in a three‐parameter model. Taking into account uncertainty in parameters such as initial population size and migration rate collectively decreased the precision of inferences dramatically. Phasing haplotypes did not improve results, regardless of sequence length. Numerous short sequences were as valuable as fewer, longer sequences, and performed best when a large sample size was sequenced at low individual depth, even when sequencing errors were added. ABC results were similar to results obtained with an alternative method based on the site frequency spectrum (SFS) when performed with unphased GBS‐type markers. We conclude that unphased GBS‐type data sets can be sufficient to precisely infer simple demographic models, and discuss possible improvements for the use of ABC with genomic data.     

Computing distribution of scale independent motifs in biological sequences

The use of Chaos Game Representation (CGR) or its generalization, Universal Sequence Maps (USM), to describe the distribution of biological sequences has been found objectionable because of the fractal structure of that coordinate system. Consequently, the investigation of distribution of symbolic motifs at multiple scales is hampered by an inexact association between distance and sequence dissimilarity. A solution to this problem could unleash the use of iterative maps as phase-state representation of sequences where its statistical properties can be conveniently investigated. In this study a family of kernel density functions is described that accommodates the fractal nature of iterative function representations of symbolic sequences and, consequently, enables the exact investigation of sequence motifs of arbitrary lengths in that scale-independent representation. Furthermore, the proposed kernel density includes both Markovian succession and currently used alignment-free sequence dissimilarity metrics as special solutions. Therefore, the fractal kernel described is in fact a generalization that provides a common framework for a diverse suite of sequence analysis techniques.     

Visual representation of DNA sequences for exon detection using non-parametric spectral estimation techniques

This paper presents a new approach for modeling of DNA sequences for the purpose of exon detection. The proposed model adopts the sum-of-sinusoids concept for the representation of DNA sequences. The objective of the modeling process is to represent the DNA sequence with few coefficients. The modeling process can be performed on the DNA signal as a whole or on a segment-by-segment basis. The created models can be used instead of the original sequences in a further spectral estimation process for exon detection. The accuracy of modeling is evaluated evaluated by using the Root Mean Square Error (RMSE) and the R-square metrics. In addition, non-parametric spectral estimation methods are used for estimating the spectral of both original and modeled DNA sequences. The results of exon detection based on original and modeled DNA sequences coincide to a great extent, which ensures the success of the proposed sum-of-sinusoids method for modeling of DNA sequences.     

Representation of DNA sequences with virtual potentials and their processing by (SEQREP) Kohonen self-organizing maps

MOTIVATION: We propose representing individual positions in DNA sequences by virtual potentials generated by other bases of the same sequence. This is a compact representation of the neighbourhood of a base. The distribution of the virtual potentials over the whole sequence can be used as a representation of the entire sequence (SEQREP code). It is a flexible code, with a length independent of the sequence size, does not require previous alignment, and is convenient for processing by neural networks or statistical techniques. RESULTS: To evaluate its biological significance, the SEQREP code was used for training Kohonen self-organizing maps (SOMs) in two applications: (a) detection of Alu sequences, and (b) classification of sequences encoding for HIV-1 envelope glycoprotein (env) into subtypes A-G. It was demonstrated that SOMs clustered sequences belonging to different classes into distinct regions. For independent test sets, very high rates of correct predictions were obtained (97% in the first application, 91% in the second). Possible areas of application of SEQREP codes include functional genomics, phylogenetic analysis, detection of repetitions, database retrieval, and automatic alignment. AVAILABILITY: Software for representing sequences by SEQREP code, and for training Kohonen SOMs is made freely available from http://www.dq.fct.unl.pt/qoa/jas/seqrep. SUPPLEMENTARY INFORMATION: Supplementary material is available at http://www.dq.fct.unl.pt/qoa/jas/seqrep/bioinf2002     

Exploration of multivariate analysis in microbial coding sequence modeling

ABSTRACT: BACKGROUND: Gene finding is a complicated procedure that encapsulates algorithms for coding sequence modeling, identification of promoter regions, issues concerning overlapping genes and more. In the present study we focus on coding sequence modeling algorithms; that is, algorithms for identification and prediction of the actual coding sequences from genomic DNA. In this respect, we promote a novel multivariate method known as Canonical Powered Partial Least Squares (CPPLS) as an alternative to the commonly used Interpolated Markov model (IMM). Comparisons between the methods were performed on DNA, codon and protein sequences with highly conserved genes taken from several species with different genomic properties. RESULTS: The multivariate CPPLS approach classified coding sequence substantially better than the commonly used IMM on the same set of sequences. We also found that the use of CPPLS with codon representation gave significantly better classification results than both IMM with protein (p < 0.001) and with DNA (p < 0.001). Further, although the mean performance was similar, the variation of CPPLS performance on codon representation was significantly smaller than for IMM (p < 0.001). CONCLUSIONS: The performance of coding sequence modeling can be substantially improved by using an algorithm based on the multivariate CPPLS method applied to codon or DNA frequencies.     

Hybridization selections, using immobilized driver DNA, to enrich for clones unique to a library

We have developed a technique which allowed us to isolate sex-specific repeats of chickens (Gallus g. domesticus) from a genomic library originally containing one sex-specific repeat clone per 300 clones. Using this plus/minus selection method, we were able to enrich a sub-library in which the sex-specific repetitive clones made up over half of the population (170-fold increase in representation). This enrichment technique used hybridization kinetics to collect clones which are bound (plus selection) or not bound (minus selection) to their respective driver DNAs immobilized on nitrocellulose paper. Plus selections enriched for repetitive sequences and removed from the library most unique sequences as well as vectors containing no inserted sequences. These plus-selected sub-libraries were enriched for repetitive clones, yet still contained sequences representing as little as 10(-4) of the genome. Minus selection removed repetitive sequences shared between the driver and the library. When a plus-selected sub-library was minus selected, the doubly selected sub-library was enriched in repetitive sequences present in the library but not the minus driver.     

A new method of global search for functional DNA segments using a fractal representation of nucleotide texts]

New global method for computer prediction of functional sites in nucleotide sequences, based on the fractal representation, is presented. Fractal representation of set of sequences (FRS) provides simple way for generating recognitions matrix of functionally similar sequences and simple estimations of its efficiency for searching homologous regions in new sequences. Other advantages of the method are absence of the necessity of sequences alignment during generating based set and searching new homologous regions and small CPU time. Usage of the method illustrated for searching globin and histone genes, for ALU repeats in human genome and long terminal repeats in virus genome.