Time delay estimation with hidden Markov models

Most traditional methods for extracting the relationships between two time series are based on cross-correlation. In a non-linear non-stationary environment, these techniques are not sufficient. We show in this paper how to use hidden Markov models to identify the lag (or delay) between different variables for such data. Adopting an information-theoretic approach, we develop a procedure for training HMMs to maximise the mutual information (MMI) between delayed time series. The method is used to model the oil drilling process. We show that cross-correlation gives no information and that the MMI approach outperforms maximum likelihood.

Bayesian online algorithms for learning in discrete Hidden Markov Models

We propose and analyze two different Bayesian online algorithms for learning in discrete Hidden Markov Models and compare their performance with the already known Baldi-Chauvin Algorithm. Using the Kullback-Leibler divergence as a measure of generalization we draw learning curves in simplified situations for these algorithms and compare their performances.

Perception of concurrent sentences with harmonic or frequency-shifted voiced excitation:performance of human listeners and of computational models based on autocorrelation

Keyword identification in one of two simultaneous sentences is improved when the sentences differ in F0, particularly when they are almost continuously voiced. Sentences of this kind were recorded, monotonised using PSOLA, and re-synthesised to give a range of harmonic ?F0s (0, 1, 3, and 10 semitones). They were additionally re-synthesised by LPC with the LPC residual frequency shifted by 25% of F0, to give excitation with inharmonic but regularly spaced components. Perceptual identification of frequency-shifted sentences showed a similar large improvement with nominal ?F0 as seen for harmonic sentences, although overall performance was about 10% poorer. We compared performance with that of two autocorrelation-based computational models comprising four stages: (i) peripheral frequency selectivity and half-wave rectification; (ii) within-channel periodicity extraction; (iii) identification of the two major peaks in the summary autocorrelation function (SACF); (iv) a template-based approach to speech recognition using dynamic time warping. One model sampled the correlogram at the target-F0 period and performed spectral matching; the other deselected channels dominated by the interferer and performed matching on the short-lag portion of the residual SACF. Both models reproduced the monotonic increase observed in human performance with increasing ?F0 for the harmonic stimuli, but not for the frequency-shifted stimuli. A revised version of the spectral-matching model, which groups patterns of periodicity that lie on a curve in the frequency-delay plane, showed a closer match to the perceptual data for frequency-shifted sentences. The results extend the range of phenomena originally attributed to harmonic processing to grouping by common spectral pattern.

The Human Genome Project: its relevance to ocular disease

This article describes the history of the Human Genome Project, how the human genome was sequenced, and analyses the likely impact which the results will have on the diagnosis, scientific understanding and, ultimately, treatment of ocular disease in the future.

Modelling data and voice traffic over IP networks using continuous-time Markov models

Common approaches to IP-traffic modelling have featured the use of stochastic models, based on the Markov property, which can be classified into black box and white box models based on the approach used for modelling traffic. White box models, are simple to understand, transparent and have a physical meaning attributed to each of the associated parameters. To exploit this key advantage, this thesis explores the use of simple classic continuous-time Markov models based on a white box approach, to model, not only the network traffic statistics but also the source behaviour with respect to the network and application. The thesis is divided into two parts: The first part focuses on the use of simple Markov and Semi-Markov traffic models, starting from the simplest two-state model moving upwards to n-state models with Poisson and non-Poisson statistics. The thesis then introduces the convenient to use, mathematically derived, Gaussian Markov models which are used to model the measured network IP traffic statistics. As one of the most significant contributions, the thesis establishes the significance of the second-order density statistics as it reveals that, in contrast to first-order density, they carry much more unique information on traffic sources and behaviour. The thesis then exploits the use of Gaussian Markov models to model these unique features and finally shows how the use of simple classic Markov models coupled with use of second-order density statistics provides an excellent tool for capturing maximum traffic detail, which in itself is the essence of good traffic modelling. The second part of the thesis, studies the ON-OFF characteristics of VoIP traffic with reference to accurate measurements of the ON and OFF periods, made from a large multi-lingual database of over 100 hours worth of VoIP call recordings. The impact of the language, prosodic structure and speech rate of the speaker on the statistics of the ON-OFF periods is analysed and relevant conclusions are presented. Finally, an ON-OFF VoIP source model with log-normal transitions is contributed as an ideal candidate to model VoIP traffic and the results of this model are compared with those of previously published work.

Progress in the development of in vitro human-based alternatives to animal models

Editorial

Further preliminary assessment of three human glioma cell lines as models of human astrocytic toxicity in vitro

Three human astroglioma lines U251-MG, U373-MG and CCF-STTG1 have been evaluated further as possible models for astrocytotoxicity (GFAP and IL-6 release). The effects of bacterial lipopolysaccharide, chloroquine diphosphate and acrylamide were studied on GFAP expression and LPS, chloroquine diphosphate, ethanol, trimethyltin chloride (TMTC) and acrylamide were examined on interleukin-6 (IL-6) release in the U373-MG line only. At 4-h LIPS elevated GFAP (17.0±5.0% P < 0.05) above control in the U251-MG cell line only. Chloroquine diphosphate over 4 h in the U251-MG line resulted in an increase in GFAP-IR to 20.3 ±4.2% and 21.1 ± 4.1 % above control levels 0.1 µM (P< 0.05) and 1 µM (P< 0.05) respectively. CQD was associated with decreases in MTT turnover, particularly after 24 h incubation. With the U373-MG line, LPS (0.5 µg/ml) increased IL-6 expression 640% above control (P < 0.001), whilst chloroquine diphosphate (100 µM), ethanol (10mM) and TMTC chloride (1 µM) also increased IL-6. It is possible that batteries of astrocytic human glioma cell lines may be applicable to the sensitive evaluation of toxicants on astrogliotic expression markers such as GFAP and IL-6.

Online learning in discrete hidden Markov models

We present and analyze three different online algorithms for learning in discrete Hidden Markov Models (HMMs) and compare their performance with the Baldi-Chauvin Algorithm. Using the Kullback-Leibler divergence as a measure of the generalization error we draw learning curves in simplified situations and compare the results. The performance for learning drifting concepts of one of the presented algorithms is analyzed and compared with the Baldi-Chauvin algorithm in the same situations. A brief discussion about learning and symmetry breaking based on our results is also presented. © 2006 American Institute of Physics.

A path-based framework for analyzing large markov models

Reliability and dependability modeling can be employed during many stages of analysis of a computing system to gain insights into its critical behaviors. To provide useful results, realistic models of systems are often necessarily large and complex. Numerical analysis of these models presents a formidable challenge because the sizes of their state-space descriptions grow exponentially in proportion to the sizes of the models. On the other hand, simulation of the models requires analysis of many trajectories in order to compute statistically correct solutions. This dissertation presents a novel framework for performing both numerical analysis and simulation. The new numerical approach computes bounds on the solutions of transient measures in large continuous-time Markov chains (CTMCs). It extends existing path-based and uniformization-based methods by identifying sets of paths that are equivalent with respect to a reward measure and related to one another via a simple structural relationship. This relationship makes it possible for the approach to explore multiple paths at the same time,· thus significantly increasing the number of paths that can be explored in a given amount of time. Furthermore, the use of a structured representation for the state space and the direct computation of the desired reward measure (without ever storing the solution vector) allow it to analyze very large models using a very small amount of storage. Often, path-based techniques must compute many paths to obtain tight bounds. In addition to presenting the basic path-based approach, we also present algorithms for computing more paths and tighter bounds quickly. One resulting approach is based on the concept of path composition whereby precomputed subpaths are composed to compute the whole paths efficiently. Another approach is based on selecting important paths (among a set of many paths) for evaluation. Many path-based techniques suffer from having to evaluate many (unimportant) paths. Evaluating the important ones helps to compute tight bounds efficiently and quickly.

Beyond the human genome: microbes, metaphors and what it means to be human in an interconnected post-genomic world

Four years after the completion of the Human Genome Project, the US National Institutes for Health launched the Human Microbiome Project on 19 December 2007. Using metaphor analysis, this article investigates reporting in English-language newspapers on advances in microbiomics from 2003 onwards, when the word “microbiome” was first used. This research was said to open up a “new frontier” and was conceived as a “second human genome project”, this time focusing on the genomes of microbes that inhabit and populate humans rather than focusing on the human genome itself. The language used by scientists and by the journalists who reported on their research employed a type of metaphorical framing that was very different from the hyperbole surrounding the decipherment of the “book of life”. Whereas during the HGP genomic successes had been mainly framed as being based on a unidirectional process of reading off information from a passive genetic or genomic entity, the language employed to discuss advances in microbiomics frames genes, genomes and life in much more active and dynamic ways.

Folate-sensitive fragile site FRA10A is due to an expansion of a CGG repeat in a novel gene, FRA10AC1, encoding a nuclear protein

Fragile sites appear visually as nonstaining gaps on chromosomes that are inducible by specific cell culture conditions. Expansion of CGG/ CCG repeats has been shown to be the molecular basis of all five folate-sensitive fragile sites characterized molecularly so far, i.e., FRAXA, FRAXE, FRAXF, FRA11B, and FRA16A. In the present study we have refined the localization of the FRA10A folate-sensitive fragile site by fluorescence in situ hybridization. Sequence analysis of a BAC clone spanning FRA10A identified a single, imperfect, but polymorphic CGG repeat that is part of a CpG island in the 5'UTR of a novel gene named FRA10ACl. The number of CGG repeats varied in the population from 8 to 13. Expansions exceeding 200 repeat units were methylated in all FRA10A fragile site carriers tested. The FRA10ACl gene consists of 19 exons and is transcribed in the centromeric direction from the FRA10A repeat. The major transcript of similar to 1450 nt is ubiquitously expressed and codes for a highly conserved protein, FRA10ACl, of unknown function. Several splice variants leading to alternative 3' ends were identified (particularly in testis). These give rise to FRA10ACl proteins with altered COOH-termini. Immunofluorescence analysis of full-length, recombinant EGFP-tagged FRA10ACl protein showed that it was present exclusively in the nucleoplasm. We show that the expression of FRA10A, in parallel to the other cloned folate-sensitive fragile sites, is caused by an expansion and subsequent methylation of an unstable CGG trinucleotide repeat. Taking advantage of three cSNPs within the FRA10ACl gene we demonstrate that one allele of the gene is not transcribed in a FRA10A carrier. Our data also suggest that in the heterozygous state FRA10A is likely a benign folate-sensitive fragile site. (C) 2004 Elsevier Inc. All rights reserved.

Transduction of CpG DNA-stimulated primary human B cells with bicistronic lentivectors.

Recently, using HIV-1-derived lentivectors, we obtained efficient transduction of primary human B lymphocytes cocultured with murine EL-4 B5 thymoma cells, but not of isolated B cells activated by CD40 ligation. Coculture with a cell line is problematic for gene therapy applications or study of gene functions. We have now found that transduction of B cells in a system using CpG DNA was comparable to that in the EL-4 B5 system. A monocistronic vector with a CMV promoter gave 32 +/- 4.7% green fluorescent protein (GFP)+ cells. A bicistronic vector, encoding IL-4 and GFP in the first and second cistrons, respectively, gave 14.2 +/- 2.1% GFP+ cells and IL-4 secretion of 1.3 +/- 0.2 ng/10(5) B cells/24 h. This was similar to results obtained in CD34+ cells using the elongation factor-1alpha promoter. Activated memory and naive B cells were transducible. After transduction with a bicistronic vector encoding a viral FLIP molecule, vFLIP was detectable by FACS or Western blot in GFP+, but not in GFP-, B cells, and 57% of sorted GFP+ B cells were protected against Fas ligand-induced cell death. This system should be useful for gene function research in primary B cells and development of gene therapies.

DNA methylation profiles of human active and inactive X chromosomes.

X-chromosome inactivation (XCI) is a dosage compensation mechanism that silences the majority of genes on one X chromosome in each female cell. To characterize epigenetic changes that accompany this process, we measured DNA methylation levels in 45,X patients carrying a single active X chromosome (X(a)), and in normal females, who carry one X(a) and one inactive X (X(i)). Methylated DNA was immunoprecipitated and hybridized to high-density oligonucleotide arrays covering the X chromosome, generating epigenetic profiles of active and inactive X chromosomes. We observed that XCI is accompanied by changes in DNA methylation specifically at CpG islands (CGIs). While the majority of CGIs show increased methylation levels on the X(i), XCI actually results in significant reductions in methylation at 7% of CGIs. Both intra- and inter-genic CGIs undergo epigenetic modification, with the biggest increase in methylation occurring at the promoters of genes silenced by XCI. In contrast, genes escaping XCI generally have low levels of promoter methylation, while genes that show inter-individual variation in silencing show intermediate increases in methylation. Thus, promoter methylation and susceptibility to XCI are correlated. We also observed a global correlation between CGI methylation and the evolutionary age of X-chromosome strata, and that genes escaping XCI show increased methylation within gene bodies. We used our epigenetic map to predict 26 novel genes escaping XCI, and searched for parent-of-origin-specific methylation differences, but found no evidence to support imprinting on the human X chromosome. Our study provides a detailed analysis of the epigenetic profile of active and inactive X chromosomes.

Epigenetic regulation of human cancer/testis antigen gene, HAGE, in chronic myeloid leukemia.

BACKGROUND AND OBJECTIVES Cancer testis antigens (CTA) provide attractive targets for cancer-specific immunotherapy. Although CTA genes are expressed in some normal tissues, such as the testis, this immunologically protected site lacks MHC I expression and as such, does not present self antigens to T cells. To date, CTA genes have been shown to be expressed in a range of solid tumors via demethylation of their promoter CpG islands, but rarely in chronic myeloid leukemia (CML) or other hematologic malignancies. DESIGN AND METHODS In this study, the methylation status of the HAGE CTA gene promoter was analyzed by quantitative methylation-specific polymerase chain reaction (MSP) and sequencing in four Philadelphia-positive cell lines (TCC-S, K562, KU812 and KYO-1) and in CML samples taken from patients in chronic phase (CP n=215) or blast crisis (BC n=47). HAGE expression was assessed by quantitative reverse transcriptase-polymerase chain reaction. RESULTS The TCC-S cell line showed demethylation of HAGE that was associated with over-expression of this gene. HAGE hypomethylation was significantly more frequent in BC (46%) than in CP (22%) (p=0.01) and was correlated with high expression levels of HAGE transcripts (p<0.0001). Of note, in CP-CML, extensive HAGE hypomethylation was associated with poorer prognosis in terms of cytogenetic response to interferon (p=0.01) or imatinib (p=0.01), molecular response to imatinib (p=0.003) and progression-free survival (p=0.05). INTERPRETATIONS AND CONCLUSION: The methylation status of the HAGE promoter directly correlates with its expression in both CML cell lines and patients and is associated with advanced disease and poor outcome.

Tissue- specific DNA methylation profiles in newborns

Experimental and epidemiological studies demonstrate that fetal growth restriction and low birth weight enhance the risk of chronic diseases in adulthood. Derangements in tissue-specific epigenetic programming of fetal and placental tissues are a suggested mechanism of which DNA methylation is best understood. DNA methylation profiles in human tissue are mostly performed in DNA from white blood cells. The objective of this study was to assess DNA methylation profiles of IGF2 DMR and H19 in DNA derived from four tissues of the newborn. We obtained from 6 newborns DNA from fetal placental tissue (n = 5), umbilical cord CD34+ hematopoietic stem cells (HSC) and CD34- mononuclear cells (MNC) (n = 6), and umbilical cord Wharton jelly (n = 5). HCS were isolated using magnetic-activated cell separation. DNA methylation of the imprinted fetal growth genes IGF2 DMR and H19 was measured in all tissues using quantitative mass spectrometry. ANOVA testing showed tissue-specific differences in DNA methylation of IGF2 DMR (p value 0.002) and H19 (p value 0.001) mainly due to a higher methylation of IGF2 DMR in Wharton jelly (mean 0.65, sd 0.14) and a lower methylation of H19 in placental tissue (mean 0.25, sd 0.02) compared to other tissues. This study demonstrates the feasibility of the assessment of differential tissue specific DNA methylation. Although the results have to be confirmed in larger sample sizes, our approach gives opportunities to investigate epigenetic profiles as underlying mechanism of associations between pregnancy exposures and outcome, and disease risks in later life.