Molecular markers of obesity and diabetes

Recently it has been shown that the consumption of a diet high in saturated fat is associated with impaired insulin sensitivity and increased incidence of type 2 diabetes. In contrast, diets that are high in monounsaturated fatty acids (MUFAs) or polyunsaturated fatty acids (PUFAs), especially very long chain n-3 fatty acids (FAs), are protective against disease. However, the molecular mechanisms by which saturated FAs induce the insulin resistance and hyperglycaemia associated with metabolic syndrome and type 2 diabetes are not clearly defined. It is possible that saturated FAs may act through alternative mechanisms compared to MUFA and PUFA to regulate of hepatic gene expression and metabolism. It is proposed that, like MUFA and PUFA, saturated FAs regulate the transcription of target genes. To test this hypothesis, hepatic gene expression analysis was undertaken in a human hepatoma cell line, Huh-7, after exposure to the saturated FA, palmitate. These experiments showed that palmitate is an effective regulator of gene expression for a wide variety of genes. A total of 162 genes were differentially expressed in response to palmitate. These changes not only affected the expression of genes related to nutrient transport and metabolism, they also extend to other cellular functions including, cytoskeletal architecture, cell growth, protein synthesis and oxidative stress response. In addition, this thesis has shown that palmitate exposure altered the expression patterns of several genes that have previously been identified in the literature as markers of risk of disease development, including CVD, hypertension, obesity and type 2 diabetes. The altered gene expression patterns associated with an increased risk of disease include apolipoprotein-B100 (apo-B100), apo-CIII, plasminogen activator inhibitor 1, insulin-like growth factor-I and insulin-like growth factor binding protein 3. This thesis reports the first observation that palmitate directly signals in cultured human hepatocytes to regulate expression of genes involved in energy metabolism as well as other important genes. Prolonged exposure to long-chain saturated FAs reduces glucose phosphorylation and glycogen synthesis in the liver. Decreased glucose metabolism leads to elevated rates of lipolysis, resulting in increased release of free FAs. Free FAs have a negative effect on insulin action on the liver, which in turn results in increased gluconeogenesis and systemic dyslipidaemia. It has been postulated that disruption of glucose transport and insulin secretion by prolonged excessive FA availability might be a non-genetic factor that has contributed to the staggering rise in prevalence of type 2 diabetes. As glucokinase (GK) is a key regulatory enzyme of hepatic glucose metabolism, changes in its activity may alter flux through the glycolytic and de novo lipogenic pathways and result in hyperglycaemia and ultimately insulin resistance. This thesis investigated the effects of saturated FA on the promoter activity of the glycolytic enzyme, GK, and various transcription factors that may influence the regulation of GK gene expression. These experiments have shown that the saturated FA, palmitate, is capable of decreasing GK promoter activity. In addition, quantitative real-time PCR has shown that palmitate incubation may also regulate GK gene expression through a known FA sensitive transcription factor, sterol regulatory element binding protein-1c (SREBP-1c), which upregulates GK transcription. To parallel the investigations into the mechanisms of FA molecular signalling, further studies of the effect of FAs on metabolic pathway flux were performed. Although certain FAs reduce SREBP-1c transcription in vitro, it is unclear whether this will result in decreased GK activity in vivo where positive effectors of SREBP-1c such as insulin are also present. Under these conditions, it is uncertain if the inhibitory effects of FAs would be overcome by insulin. The effects of a combination of FAs, insulin and glucose on glucose phosphorylation and metabolism in cultured primary rat hepatocytes at concentrations that mimic those in the portal circulation after a meal was examined. It was found that total GK activity was unaffected by an increased concentration of insulin, but palmitate and eicosapentaenoic acid significantly lowered total GK activity in the presence of insulin. Despite the fact that total GK enzyme activity was reduced in response to FA incubation, GK enzyme translocation from the inactive, nuclear bound, to active, cytoplasmic state was unaffected. Interestingly, none of the FAs tested inhibited glucose phosphorylation or the rate of glycolysis when insulin is present. These results suggest that in the presence of insulin the levels of the active, unbound cytoplasmic GK are sufficient to buffer a slight decrease in GK enzyme activity and decreased promoter activity caused by FA exposure. Although a high fat diet has been associated with impaired hepatic glucose metabolism, there is no evidence from this thesis that FAs themselves directly modulate flux through the glycolytic pathway in isolated primary hepatocytes when insulin is also present. Therefore, although FA affected expression of a wide range of genes, including GK, this did not affect glycolytic flux in the presence of insulin. However, it may be possible that a saturated FA-induced decrease in GK enzyme activity when combined with the onset of insulin resistance may promote the dys-regulation of glucose homeostasis and the subsequent development of hyperglycaemia, metabolic syndrome and type 2 diabetes.

Towards improved speech recognition for resource poor languages

In recent times, the improved levels of accuracy obtained by Automatic Speech Recognition (ASR) technology has made it viable for use in a number of commercial products. Unfortunately, these types of applications are limited to only a few of the world’s languages, primarily because ASR development is reliant on the availability of large amounts of language specific resources. This motivates the need for techniques which reduce this language-specific, resource dependency. Ideally, these approaches should generalise across languages, thereby providing scope for rapid creation of ASR capabilities for resource poor languages. Cross Lingual ASR emerges as a means for addressing this need. Underpinning this approach is the observation that sound production is largely influenced by the physiological construction of the vocal tract, and accordingly, is human, and not language specific. As a result, a common inventory of sounds exists across languages; a property which is exploitable, as sounds from a resource poor, target language can be recognised using models trained on resource rich, source languages. One of the initial impediments to the commercial uptake of ASR technology was its fragility in more challenging environments, such as conversational telephone speech. Subsequent improvements in these environments has gained consumer confidence. Pragmatically, if cross lingual techniques are to considered a viable alternative when resources are limited, they need to perform under the same types of conditions. Accordingly, this thesis evaluates cross lingual techniques using two speech environments; clean read speech and conversational telephone speech. Languages used in evaluations are German, Mandarin, Japanese and Spanish. Results highlight that previously proposed approaches provide respectable results for simpler environments such as read speech, but degrade significantly when in the more taxing conversational environment. Two separate approaches for addressing this degradation are proposed. The first is based on deriving better target language lexical representation, in terms of the source language model set. The second, and ultimately more successful approach, focuses on improving the classification accuracy of context-dependent (CD) models, by catering for the adverse influence of languages specific phonotactic properties. Whilst the primary research goal in this thesis is directed towards improving cross lingual techniques, the catalyst for investigating its use was based on expressed interest from several organisations for an Indonesian ASR capability. In Indonesia alone, there are over 200 million speakers of some Malay variant, provides further impetus and commercial justification for speech related research on this language. Unfortunately, at the beginning of the candidature, limited research had been conducted on the Indonesian language in the field of speech science, and virtually no resources existed. This thesis details the investigative and development work dedicated towards obtaining an ASR system with a 10000 word recognition vocabulary for the Indonesian language.

Automatic speaker recognition under adverse conditions

Speaker verification is the process of verifying the identity of a person by analysing their speech. There are several important applications for automatic speaker verification (ASV) technology including suspect identification, tracking terrorists and detecting a person’s presence at a remote location in the surveillance domain, as well as person authentication for phone banking and credit card transactions in the private sector. Telephones and telephony networks provide a natural medium for these applications. The aim of this work is to improve the usefulness of ASV technology for practical applications in the presence of adverse conditions. In a telephony environment, background noise, handset mismatch, channel distortions, room acoustics and restrictions on the available testing and training data are common sources of errors for ASV systems. Two research themes were pursued to overcome these adverse conditions: Modelling mismatch and modelling uncertainty. To directly address the performance degradation incurred through mismatched conditions it was proposed to directly model this mismatch. Feature mapping was evaluated for combating handset mismatch and was extended through the use of a blind clustering algorithm to remove the need for accurate handset labels for the training data. Mismatch modelling was then generalised by explicitly modelling the session conditions as a constrained offset of the speaker model means. This session variability modelling approach enabled the modelling of arbitrary sources of mismatch, including handset type, and halved the error rates in many cases. Methods to model the uncertainty in speaker model estimates and verification scores were developed to address the difficulties of limited training and testing data. The Bayes factor was introduced to account for the uncertainty of the speaker model estimates in testing by applying Bayesian theory to the verification criterion, with improved performance in matched conditions. Modelling the uncertainty in the verification score itself met with significant success. Estimating a confidence interval for the "true" verification score enabled an order of magnitude reduction in the average quantity of speech required to make a confident verification decision based on a threshold. The confidence measures developed in this work may also have significant applications for forensic speaker verification tasks.