Day 1 :
Keynote Forum
Tao Huang
Shanghai Institutes for Biological Sciences, China
Keynote: Deciphering the mechanisms of complex diseases using network and machine learning approaches
Time : 09:40-10:25
Biography:
Tao Huang has completed his Postdoctoral Fellowship in the Department of Genetics and Genomics Sciences, Icahn School of Medicine at Mount Sinai, New York City, USA. He is currently a Professor and the Director of Bioinformatics Core Facility at the Institute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences ( P R China). His research interests include bioinformatics, computational biology, systems genetics and big data research. He has published over 100 articles in peer-reviewed journals. His works have been cited for 3138 times with an h-index of 26 and an i10-index of 64. He has been an Independent Reviewer for 30 journals. He is the Editor of the book: “Computational Systems Biology - Methods and Protocols” published in Springer; Guest Editor for BBA Molecular Basis of Disease, BBA General Subjects, Artificial Intelligence in Medicine, BioMed Research International, Combinatorial Chemistry & High Throughput Screening and Computational and Mathematical Methods in Medicine.
Abstract:
The complex diseases, such as cancer, involves dysfunctions on multiple levels. There have been several methods to identify the dysfunctions on each level. On genomics level, GWAS (Genome-wide association study) can detect the disease phenotype associated SNPs (single nucleotide polymorphisms). But many GWAS identified SNPs locate in intergenic region and can't be annotated to specific genes. With the help of eQTL (expression quantitative trait loci) method, the downstream genes of the GWAS SNPs can be found. And based on the co-expression network, regulatory network, protein-protein interaction network, or even protein-chemical interaction network of these genes, the possible cascade of the signal transduction from the genetic perturbations can be discovered. By integrating all these omics data, we can establish a systems biology model of how the GWAS SNPs affect the expression of their direct target genes, how these direct targets affect secondary target genes, or proteins, or metabolites, and eventually cause the catastrophic pathological changes on network or pathway. With the comprehensive network, we can not only study the mechanism of one disease, but also the relationship of several traits. For example, schizophrenia and anti-tuberculosis drug-induced hepatotoxicity co-occurred frequently but they do not share common genes. We mapped their genes onto the network, the common drivers of these two diseases were revealed through two-way RWR analysis. Another example is that the lung cancer dysfunctions on several levels but how the different level dysfunctions connect are still unknown. We investigated the mutation, methylation, mRNA and microRNA expression difference between cancer and normal tissues using machine learning methods, then mapped the dysfunction of each level onto the comprehensive network. By connecting each other with shortest paths, the highly frequent shortest path genes that can cause the tumorigenesis on multiple levels were discovered. Overall, the network and machine learning approaches are effective to dissect the complex system and reveal the mechanisms of diseases.
Break: Networking & Refreshments 10:25-10:45 @ Europa Foyer
Keynote Forum
Andrea Ragusa
University of Salento, Italy
Keynote: Metabolomics and lipidomic profile of clinical serum samples from patients with liver disease
Time : 10:45-11:30
Biography:
Andrea Ragusa pursued his PhD from the University of Southampton (UK) in 2005. He dedicated himself to nanomedicine and metabolomics by developing novel smart nanovectors for the targeted delivery of drugs and studying their effects in vitro through NMR and HPLC-MS investigations. He is an Organic Chemist with an expertise in analytical techniques. More recently, he shifted his efforts towards the analysis of clinical samples for better understanding the molecular mechanisms involved in several liver-related diseases, with particular attention to HCC (Hepatocellular carcinoma), as well as for identifying novel biomarkers and their exploitation, in combination with multivariate statistical analysis, in the early diagnosis and treatment of the disease.
Abstract:
Statement of the Problem: Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and the third leading cause of cancer deaths worldwide. Unfortunately, early-stage HCC is asymptomatic and most patients already have an advanced stage at the time of diagnosis. Alpha fetal protein (AFP) is the current clinical biomarker used for the diagnosis and follow-up of HCC treatments. Nevertheless, its sensitivity is not satisfactory and there is an urgent need for the identification of novel more efficient biomarkers. In this regard, metabolomics can provide unique clinical information and it offers the possibility to understand the system-level effects of metabolites through its integration into multi-functional approaches. Nuclear magnetic resonance (NMR) spectroscopy already showed to be a valuable tool for metabolomic investigations, providing a system capable of studying and identifying novel biomarkers, as well as for yielding a metabolic fingerprint useful for the classification of several pathologies.
Methodology: 1D and 2D 1H NMR spectroscopy, in combination with univariate and multivariate statistical analysis (MVA), was used to investigate the metabolic profile of serum samples from HCC, type-2 diabetes mellitus (T2DM), and control subjects. The patients’ response to metformin or insulin treatment was also investigated. A similar analysis was also performed on the corresponding serum lipid extracts, yielding a more global view of the metabolic alterations occurring.
Findings: Several metabolites directly or indirectly related to energy pathways were found to be either over- or under-expressed in pathological patients with respect to control patients. Among them, a substantial increase in lactate production, with the concomitant decrease of branched amino acids, was observed. Similarly, an increase in the total fatty acids (FA) with a concomitant decrease of unsaturated FA, phosphocholines, phospholipids, and cholesterol was also detected.
Conclusion & Significance: NMR spectroscopy, in combination with MVA, allowed to gain a better insight into the metabolic alterations involved in HCC disease. The obtained metabolic fingerprints could be exploited in the early diagnosis of the disease
as well as in its follow-up upon treatment.
Keynote Forum
Gerald C Hsu
eclaireMD, USA
Keynote: Relationship between metabolism and obesity along with type 2 diabetes and cardiovascular risk
Time : 11:30-12:15
Biography:
Gerald C Hsu received an honorable PhD in mathematics and majored in engineering at MIT. He attended different universities over 17 years and studied seven academic disciplines. He has spent 20,000 hours in T2D research. First, he studied six metabolic diseases and food nutrition during 2010-2013, then conducted research during 2014-2018. His approach is “math-physics and quantitative medicine” based on mathematics, physics, engineering modeling, signal processing, computer science, big data analytics, statistics, machine learning, and AI. His main focus is on preventive medicine using prediction tools. He believes that the better the prediction, the more control you have. The author has not received any financial assistance from any organization.
Abstract:
Introduction: By using the big data on one patient (author), this clinical paper describes the relationship between the metabolism state and medical conditions, including obesity, diabetes, and cardiovascular risk.
Methodology: The obese patient was diagnosed with type 2 diabetes (T2D), hyperlipidemia, hypertension for over 25 years along with suffering five cardiac episodes during 1994-2006. The study processed ~1.5M detailed metabolic conditions and lifestyle data (2012-2018) based on a math-physical medicine approach (mathematics, physics, engineering modeling, artificial intelligence or AI), rather than the traditional biochemical medicine method. The author defined two new terms: Metabolism Index (MI) and General Health Status Unit (GHSU) to evaluate a person’s overall metabolism and associated chronic diseases. The research steps are 1. observing the patient’s physical phenomena and metabolic changes, collecting relevant big data. 2. building up engineering models and deriving inter-relationship equations, applying statistics tools for variance study. 3. using machine learning and AI to predict important metabolic changes.
Results: Metabolic factors and health symptoms show some key results within eight years.
Conclusions: This math-physical medicine approach has proven the close relationship between metabolic changes quantitatively
due to the improvement of lifestyle management and chronic disease conditions.
- Computational Methodologies | Systems Biology | Metabolomics
Location: Athens
Chair
Tao Huang
Shanghai Institutes for Biological Sciences, China
Co-Chair
Andrea Ragusa
University of Salento, Italy
Session Introduction
Abdullah Basoglu
Selcuk University, Turkey
Title: Effects of boron supplementation on peripartum dairy cows’ health: A metabolomic approach
Time : 12:15-12:45
Biography:
Abdullah Basoglu, Prof.Dr has been working at Selcuk University, as an expert on NMR based matabolomics, boron metabolism and metabolic diseases seen in animals.He coordinated some proposals related to metabolomics at University of Florence, CERM in Italy.
Abstract:
Although many different dietary studies on the prevention of negative energy balance-related diseases are often encountered, this is the first study investigating the effects of boron supplementation on peripartum dairy cows’ health in the light of an omics approach. Twenty-eight healthy cows (1 control and 3 experimental groups) were enrolled from 2 months before predicted calving until 2 months after calving. Experimental groups were assigned to receive boron at increasing doses as an oral bolus. Production parameters, biochemical profile, Nuclear magnetic resonance-based metabolomics profile, and mRNA abundance of gluconeogenic enzymes and lipid oxidation genes were determined. Pivotal knowledge was obtained on boron distribution in the body. Production parameters and mRNA abundance of the genes were not affected by the treatments. Postpartum nonesterified fatty acids, β-hydroxybutyrate, and triglyceride concentrations were significantly decreased in experimentals. The primary differences among groups were in lipid-soluble metabolites. There were significant differences in metabolites including postpartum valine, β-hydroxybutyrate, polyunsaturated fatty acid and citrate, propionate, isobutyrate, choline metabolites (betaine, phosphatidylcholine, and sphingomyelin) and some types of fatty acids and cholesterol in experimentals. Boron appears to be effective in minimizing negative energy balance and improving health of postpartum dairy cows.
Alla Karnovsky
University of Michigan, USA
Title: Tools for data-driven network analysis of metabolomics data
Time : 12:45 - 13:15
Biography:
Alla Karnovsky pursued her PhD in Cell and Developmental Biology from the Russian Academy of Sciences, Russia. She did her Postdoctoral work at the University of Colorado Boulder, USA followed by nine years of bioinformatics work in pharmaceutical industry. Currently, she is a Research Assistant Professor of Computational Medicine and Bioinformatics at the University of Michigan, USA. Her research interests involve the analysis of high throughput omics data, focusing primarily on metabolomics, and the development of computational methods and tools for the analysis and integration of metabolomics data with other types of genomic data.
Abstract:
Metabolomics has established itself as a powerful platform to interrogate cellular biochemistry, identify novel biomarkers and provide insights into biochemical mechanisms of disease. As metabolomics data sets become increasingly large and complex, there is a growing need for data analysis and visualization tools to help interpret experimentally observed changes and put them into relevant biological context. A common approach to interpreting the results of metabolomics and lipidomics experiments is to map and visualize experimentally measured metabolites in the context of known biochemical pathways. Several tools for performing this type of analysis have been developed including our own tool Metscape. Some tools have adopted Functional Enrichment Testing methods developed for gene expression data for the analysis of metabolomics data. However, the scope of their application has been limited to known compounds from large, well-annotated pathways, which are often occupied by a small portion of the experimentally measured metabolome. An alternative to knowledge-based data analysis is to infer meaningful associations between metabolites/lipids from experimental data and build data-driven metabolic networks to help generate biological insights. We developed a new Differential Network Enrichment Analysis method (DNEA) that uses joint structural sparsity estimation to build partial correlation networks from the data (for two or more experimental conditions), performs consensus clustering to identify highly connected network components (subnetworks) and uses Network-based Gene Set Analysis (NetGSA) to identify the differentially enriched subnetworks. We will present the applications of DNEA for the analysis of metabolomics and lipidomics data and demonstrate that it allows identifying alterations in both network structure and expression levels of interacting biomolecules that impact disease phenotypes.
Break: Lunch Break 13:20-14:20 @ La Place AB
Yan Ji
Shanghai Institutes for Biological Sciences, China
Title: SR splicing, splicing-ratio based splicing detection method
Time : 14:20-14:50
Biography:
Yan Ji obtained his PhD Degree in Bioinformatics from Shanghai Institutes for Biological Science of the Chinese Academy of Sciences, P R China. Before May 2015, he worked as a Translational Researcher at the Innovation Center of China, AstraZeneca (a pharmaceutical company). Since June 2015, he has been an Assistant Professor in Shanghai Institutes for Biological Science of the Chinese Academy of Sciences, P R China and studied some tumour types which are prevalent in China by using high-throughput data analysis from different perspectives including splicing and long non-coding RNAs. He has published more than 8 papers in reputed journals. In this conference, he will present a novel method detecting splicing events which has been successfully used to discover PTEN-regulated splicing events.
Abstract:
Through alternative splicing, most human genes express multiple isoforms that often differ in function. To infer isoform regulation from high-throughput sequencing of cDNA fragments (RNA-seq), a pipeline is developed to detect differential alternative splicing events between RNA-Seq samples of treatment and control conditions. It is based on a metric defined as splicing ratio (SR) which was used in some splicing-related studies (SR Splicing is flexible to handle different types of study design. Either control group or treatment group can have single or replicated/pooled samples. It can analyze all major types
of alternative splicing patterns and use RNA-Seq reads and use RNA-Seq reads mapped to splice junctions. By comparing with other commonly used software, the performance of SR Splicing is evaluated from three aspects: sensitivity, accuracy, and validation. SR Splicing has comparable sensitivity and higher accuracy. Experimental validation using qRT-PCR (quantitative real time polymerase chain reaction) confirmed a selected set of splicing events that are significantly changed in Pten knock out data of the mouse MEF cell line, demonstrating the utility of the approach applied on experimental biological data sets.SR Splicing used different statistics and less filtering to return a list of significantly changed splicing events, with associated p values and false discovery corrections. It includes detailed information on the detected splicing differences such as which exon/ junctions are involved, alternative splice type (skipped exon, mutually exclusive exons, retained intron, alternative 5' splice site, and alternative 3' splice site), magnitude of difference, and coverage.
Maroun Bou Sleiman
Swiss Federal Institute of Technology in Lausanne, Switzerland
Title: Enabling discovery through multi-omics and systems genetics
Time : 14:50-15:20
Biography:
Maroun Bou Sleiman obtained his PhD in the Laboratories of Professor Bart Deplancke and Professor Bruno Lemaitre, working on the genetic and molecular bases of infection in the Drosophila Genetic Reference Panel. He is a Scientist in the Laboratory of Integrative Systems Physiology headed by Professor Johan Auwerx at the Swiss Federal Institute of Technology (EPFL) in Lausanne (Switzerland). He works at the frontier between genetics, systems biology, and multi-omics in order to understand complex traits and diseases in different panels such as the BXD, HDP, and ITP in mice, as well as Caenorhabditis elegans RIAILs (recombinant inbred advanced intercross lines). These systems genetics approaches pave the way for a better mechanistic understanding of complex biological systems and pave the way for mechanistically-informed precision medicine.
Abstract:
Our laboratory is using systems approaches to map the signaling networks that govern mitochondrial function and as such regulate organismal metabolism in health, aging and disease. We apply a state-of-the-art biological toolkit to study a variety of model systems, ranging from the plant Arabidopsis thaliana, over the nematode Caenorhabditis elegans, to the mouse and all the way to humans. Our research has not only allowed the development of new methodologies and scientific approaches applied to population, as exemplified by the development of cross-species multi-layered genetics/omics gene mapping strategies, but also contributed to improved understanding of how signaling pathways control mitochondrial function and metabolism. Although our research addresses basic biomedical questions, we aim at translating our research into novel preventive and therapeutic strategies for common diseases, such as type 2 diabetes, frailty, and obesity, as well as rare inherited mitochondrial diseases. The translational value of our work is testified by the fact that several drugs targeting processes and pathways which we elucidated are currently used in the clinic.
Ryohei Eguchi
Nara Institute of Science and Technology, Japan
Title: Classification of biosynthesis pathways of alkaloids using graph convolutional neural networks
Time : 15:20-15:50
Biography:
Ryohei Eguchi obtained his Doctor’s Degree in Engineering at the Graduate School of Information Science of Nara Institute of Science and Technology, Japan
(2018) and in the same year joined the institute as a Doctoral Researcher at the Data Driven Science Creation Center. He is involved in the research on gene expression analysis and metabolic pathway prediction of secondary metabolites by applying multivariate analysis and machine learning techniques, mainly in the
field of bioinformatics and chemoinformatics.
Abstract:
Secondary metabolites are defined as a group of natural compounds that are not directly involved in growth, development or reproduction of organisms. The term ‘secondary’ in the context of metabolic pathways, was introduced in 1891. The functions of secondary metabolites are related to their prevalent potent biomolecular activity acquired in evolution process involving pest and pathogen defense, UV-B-sunscreens etc. The classification of secondary metabolites taking chemical structure and metabolic pathway into consideration could provide important clues to activities of metabolites which lead to interpretation of function acquisition mechanisms of secondary metabolites in evolutional process. And, various chemical descriptors such like molecular fingerprints have been long discussed to represent biochemical features, in order to embed molecular structures into a numerical space and quantify their activities. However, it is still difficult to predict bioactivities from molecular structures since it depends on the choices of those chemical descriptors. Recently, machine learning methods based on Graph Convolutional Neural Networks (GCNN) have been proposed that can automatically optimize a model for molecular feature extraction from the given training sets. This study introduces an application of GCNN to predict metabolic pathways of alkaloids, namely, one of the largest family of secondary metabolites in plants. The author trained and tested GCNN model on 578 alkaloid compounds and the mean accuracy of 20 runs with random sampling is about 94% (Number of epoch: 200). It is greatly expected that it will lead to an understanding of the evolution of metabolic system unique to organisms.
Break: Networking & Refreshments 15:50-16:10 @ Europa Foyer