Julia Tzu-Ya Weng

Myocardial infarction classification with multi-lead ECG using hidden Markov models and Gaussian mixture models

This study presented a new diagnosis system for myocardial infarction classification by converting multi-lead ECG data into a density model for increasing accuracy and flexibility of diseases detection. In contrast to the traditional approaches, a hybrid system with HMMs and GMMs was employed for data classification. A hybrid approach using multi-leads, i.e., lead-V1, V2, V3 and V4 for myocardial infarction were developed and HMMs were used not only to find the ECG segmentations but also to calculate the log-likelihood value which was treated as statistical feature data of each heartbeats ECG complex. The 4-dimension feature vector extracted by HMMs was clustered by GMMs with different numbers of distribution (disease and normal data). SVMs classifier was also examined for comparison with our system in experimental result. There were total 1129 samples of heartbeats from clinical data, including 582 data with myocardial infarction and 547 normal data. The sensitivity of this diagnosis system achieved 85.71%, specificity achieved 79.82% and accuracy achieved 82.50% statistically.

Integrative epigenetic profiling analysis identifies DNA methylation changes associated with chronic alcohol consumption

Alcoholism has always been a major public health concern in Taiwan, especially in the aboriginal communities. Emerging evidence supports the association between DNA methylation and alcoholism, though very few studies have examined the effect of chronic alcohol consumption on the epignome. Since 1986, we have been following up on the mental health conditions of four major aboriginal peoples of Taiwan. The 993 aboriginal people who underwent the phase 1 (1986) clinical interviews were followed up through phase 2 (1990-1992), and phase 3 (2003-2009). Selected individuals for the current study included 10 males from the phase 1 normal cohort who remained normal at phase 2 and became dependent on alcohol by phase 3 and 10 control subjects who have not had any drinking problems throughout the study. We profiled the DNA methylation changes in the blood samples collected at phases 2 and 3. Enrichment analyses have identified several biological processes related to immune system responses and aging in the control group. In contrast, differentially methylated genes in the case group were mostly associated with susceptibility to infections, as well as pathways related to muscular contraction and neural degeneration. The methylation levels of six genes were found to correlate with alcohol consumption. These include genes involved in neurogenesis (NPDC1) and inflammation (HERC5), as well as alcoholism-associated genes ADCY9, CKM, and PHOX2A. Given the limited sample size, our approach uncovered genes and disease pathways associated with chronic alcohol consumption at the epigenetic level. The results offer a preliminary methylome map that enhances our understanding of alcohol-induced damages and offers new targets for alcohol injury research.

An interpretable rule-based diagnostic classification of diabetic nephropathy among type 2 diabetes patients

BackgroundThe prevalence of type 2 diabetes is increasing at an alarming rate. Various complications are associated with type 2 diabetes, with diabetic nephropathy being the leading cause of renal failure among diabetics. Often, when patients are diagnosed with diabetic nephropathy, their renal functions have already been significantly damaged. Therefore, a risk prediction tool may be beneficial for the implementation of early treatment and prevention.ResultsIn the present study, we developed a decision tree-based model integrating genetic and clinical features in a gender-specific classification for the identification of diabetic nephropathy among type 2 diabetic patients. Clinical and genotyping data were obtained from a previous genetic association study involving 345 type 2 diabetic patients (185 with diabetic nephropathy and 160 without diabetic nephropathy). Using a five-fold cross-validation approach, the performance of using clinical or genetic features alone in various classifiers (decision tree, random forest, Naïve Bayes, and support vector machine) was compared with that of utilizing a combination of attributes. The inclusion of genetic features and the implementation of an additional gender-based rule yielded better classification results.ConclusionsThe current model supports the notion that genes and gender are contributing factors of diabetic nephropathy. Further refinement of the proposed approach has the potential to facilitate the early identification of diabetic nephropathy and the development of more efficient treatment in a clinical setting.

Gene expression profiling of biological pathway alterations by radiation exposure.

Though damage caused by radiation has been the focus of rigorous research, the mechanisms through which radiation exerts harmful effects on cells are complex and not well-understood. In particular, the influence of low dose radiation exposure on the regulation of genes and pathways remains unclear. In an attempt to investigate the molecular alterations induced by varying doses of radiation, a genome-wide expression analysis was conducted. Peripheral blood mononuclear cells were collected from five participants and each sample was subjected to 0.5 Gy, 1 Gy, 2.5 Gy, and 5 Gy of cobalt 60 radiation, followed by array-based expression profiling. Gene set enrichment analysis indicated that the immune system and cancer development pathways appeared to be the major affected targets by radiation exposure. Therefore, 1 Gy radioactive exposure seemed to be a critical threshold dosage. In fact, after 1 Gy radiation exposure, expression levels of several genes including FADD, TNFRSF10B, TNFRSF8, TNFRSF10A, TNFSF10, TNFSF8, CASP1, and CASP4 that are associated with carcinogenesis and metabolic disorders showed significant alterations. Our results suggest that exposure to low-dose radiation may elicit changes in metabolic and immune pathways, potentially increasing the risk of immune dysfunctions and metabolic disorders.

Systematic Expression Profiling Analysis Identifies Specific MicroRNA-Gene Interactions that May Differentiate between Active and Latent Tuberculosis Infection

Tuberculosis (TB) is the second most common cause of death from infectious diseases. About 90% of those infected are asymptomatic—the so-called latent TB infections (LTBI), with a 10% lifetime chance of progressing to active TB. To further understand the molecular pathogenesis of TB, several molecular studies have attempted to compare the expression profiles between healthy controls and active TB or LTBI patients. However, the results vary due to diverse genetic backgrounds and study designs and the inherent complexity of the disease process. Thus, developing a sensitive and efficient method for the detection of LTBI is both crucial and challenging. For the present study, we performed a systematic analysis of the gene and microRNA profiles of healthy individuals versus those affected with TB or LTBI. Combined with a series of in silico analysis utilizing publicly available microRNA knowledge bases and published literature data, we have uncovered several microRNA-gene interactions that specifically target both the blood and lungs. Some of these molecular interactions are novel and may serve as potential biomarkers of TB and LTBI, facilitating the development for a more sensitive, efficient, and cost-effective diagnostic assay for TB and LTBI for the Taiwanese population.

A novel algorithm for mining closed temporal patterns from interval-based data

Closed sequential patterns have attracted researchers’ attention due to their capability of using compact results to preserve the same expressive power as conventional sequential patterns. However, studies to date have mainly focused on mining conventional patterns from time interval-based data, where each datum persists for a period of time. Few research efforts have elaborated on discovering closed interval-based sequential patterns (also referred to as closed temporal patterns). Mining closed temporal patterns are an arduous problem since the pairwise relationships between two interval-based events are intrinsically complex. In this paper, we develop an efficient algorithm, CCMiner, which stands for Closed Coincidence Miner to discover frequent closed patterns from interval-based data. The algorithm also employs some optimization techniques to effectively reduce the search space. The experimental results on both synthetic and real datasets indicate that CCMiner not only significantly outperforms the prior interval-based mining algorithms in execution time but also possesses graceful scalability. Furthermore, we also apply CCMiner to a real dataset to show the practicability of time interval-based closed pattern mining.

UbiNet: an online resource for exploring the functional associations and regulatory networks of protein ubiquitylation

Protein ubiquitylation catalyzed by E3 ubiquitin ligases are crucial in the regulation of many cellular processes. Owing to the high throughput of mass spectrometry-based proteomics, a number of methods have been developed for the experimental determination of ubiquitylation sites, leading to a large collection of ubiquitylation data. However, there exist no resources for the exploration of E3-ligase-associated regulatory networks of for ubiquitylated proteins in humans. Therefore, the UbiNet database was developed to provide a full investigation of protein ubiquitylation networks by incorporating experimentally verified E3 ligases, ubiquitylated substrates and protein–protein interactions (PPIs). To date, UbiNet has accumulated 43 948 experimentally verified ubiquitylation sites from 14 692 ubiquitylated proteins of humans. Additionally, we have manually curated 499 E3 ligases as well as two E1 activating and 46 E2 conjugating enzymes. To delineate the regulatory networks among E3 ligases and ubiquitylated proteins, a total of 430 530 PPIs were integrated into UbiNet for the exploration of ubiquitylation networks with an interactive network viewer. A case study demonstrated that UbiNet was able to decipher a scheme for the ubiquitylation of tumor proteins p63 and p73 that is consistent with their functions. Although the essential role of Mdm2 in p53 regulation is well studied, UbiNet revealed that Mdm2 and additional E3 ligases might be implicated in the regulation of other tumor proteins by protein ubiquitylation. Moreover, UbiNet could identify potential substrates for a specific E3 ligase based on PPIs and substrate motifs. With limited knowledge about the mechanisms through which ubiquitylated proteins are regulated by E3 ligases, UbiNet offers users an effective means for conducting preliminary analyses of protein ubiquitylation. The UbiNet database is now freely accessible via http://csb.cse.yzu.edu.tw/UbiNet/. The content is regularly updated with the literature and newly released data. Database URL: http://csb.cse.yzu.edu.tw/UbiNet/.

Metagenome and Metatranscriptome Profiling of Moderate and Severe COPD Sputum in Taiwanese Han Males.

Chronic obstructive pulmonary disease (COPD) is an inflammatory lung disorder characterized by the progressive obstruction of airflow and is currently the fourth leading cause of death in the world. The pathogenesis of COPD is thought to involve bacterial infections and inflammations. Owing to advancement in sequencing technology, evidence is emerging that supports an association between the lung microbiome and COPD. However, few studies have looked into the expression profile of the bacterial communities in the COPD lungs. In this study, we analyzed the sputum microbiome of four moderate and four severe COPD male patients both at the DNA and RNA level, using next generation sequencing technology. We found that bacterial composition determined by 16S rRNA gene sequencing may not directly translate to the set of actively expressing bacteria as defined by transcriptome sequencing. The two sequencing data agreed on Prevotella, Rothia, Neisseria, Porphyromonas, Veillonella, Fusobacterium and Streptococcus being among the most differentially abundant genera between the moderate and severe COPD samples, supporting their association with COPD severity. However, the two sequencing analyses disagreed on the relative abundance of these bacteria in the two COPD groups, implicating the importance of studying the actively expressing bacteria for enriching our understanding of COPD. Though we have described the metatranscriptome profiles of the lung microbiome in moderate and severe COPD, further investigations are required to determine the functional basis underlying the relationship between the microbial species in the lungs and pathogenesis of COPD.

Myocardial Infarction Classification by Morphological Feature Extraction from Big 12-Lead ECG Data

Rapid and accurate diagnosis of patients with acute myocardial infarction is vital. The ST segment in Electrocardiography (ECG) represents the change of electric potential during the period from the end of ventricular depolarization to the beginning of repolarization and plays an important role in the detection of myocardial infarction. However, ECG monitoring generates big volumes of data and the underlying complexity must be extracted by a combination of methods. This study combines the advantages of polynomial approximation and principal component analysis. The proposed approach is stable for the 12-lead ECG data collected from the PTB database and achieves an accuracy of 98.07 %.

Incrementally mining temporal patterns in interval-based databases

In several real-life applications, sequence databases, in general, are updated incrementally with time. Some discovered sequential patterns may be invalidated and some new ones may be introduced by the evolution of the database. When a small set of sequences grow, or when some new sequences are added into the database, re-mining sequential patterns from scratch each time is usually inefficient and thus not feasible. Although there have been several recent studies on the maintenance of sequential patterns in an incremental manner, these works only consider the patterns extracted from time point-based data. Few research efforts have been elaborated on maintaining time interval-based sequential patterns, also called temporal patterns, where each datum persists for a period of time. In this paper, an efficient algorithm, Inc_TPMiner (Incremental Temporal Pattern Miner) is developed to incrementally discover temporal patterns from interval-based data. Moreover, the algorithm employs some optimization techniques to reduce the search space effectively. The experimental results on both synthetic and real datasets indicate that Inc_TPMiner significantly outperforms re-mining with static algorithms in execution time and possesses graceful scalability. Furthermore, we also apply Inc_TPMiner on a real dataset to show the practicability of incremental mining of temporal patterns.