The 4th Big Data Forum for Life and Health Sciences

The human microbiome, that is the totality of microbes inhabiting us, is spread over various body sites, the largest and probably medically most relevant one is the human gut. The gut microbiome is a complex system, which is still little understood and it is not even clear how the healthy state(s) should be defined. Yet, an increasing number of diseases is being associated with dysbiosis of the microbiome, often implied by metagenome-wide association studies (MWAS). MWAS do not reveal causalities, but disease associations and thus can still be a starting point for diagnostics, as I will illustrate using colorectal cancer (Wirbel et al., Nature Med., 2019). MWAS results are also often indirect or confounded (Schmidt et al, Cell 2018), the latter we could demonstrate for type 2 diabetes, where the intake of the drug metformin rather than the disease itself leads to the association with the gut microbiome (Forslund et al., Nature 2015). In an in vitro screen of 1200 marketed drugs against each of 40 human gut bacteria we could show that a quarter of all non-antibiotic drugs directly inhibit at least one gut microbial strain (Maier et al., Nature, 2018), a taxonomic resolution level that is becoming also achievable in MWAS (Schloissnig et al. 2013). Using single nucleotide variants (SNVs) we are able, for example, to trace donor and recipient SNVs after faecal microbiota transplantation, a microbial therapy (Li et al., Science 2016), or to trace oral-gut transmission (Schmidt et al, elife 2019). As our microbes are coming from the environment and our planet is one big ecosystem, it is crucial to study biodiversity and interactions of microbes at planetary scale. The feasibility of such a global approach is illustrated by (i) the TARA oceans project, surveying the microbial diversity of this vast ecosystem by studying plankton in all major ocean regions (Bork et al., Science 2015 and refs therein) and (ii) topsoil microbiomics (Bahram et al., Nature, 2018), revealing a global war between fungi and bacteria as well as regional antibiotic resistance gene reservoirs.

The widely applied single cell genomic technology accelerates the researches on cell heterogeneity at the levels of DNA, gene expression profiles and chromatin topology by labeling individual cells with cellular barcode. We applied this technology in analyzing the endometrium tissue from Recurrent Implantation Failure (RIF) patients and healthy controls and characterized the distinct cell compositions between them. We demonstrated that the diversity of stromal cell of RIF patients is significantly lower than healthy control, which is consistent with their differences in endometrial lining thickness. We also detected a gene expression pattern that matches mid-to-late luteal phase in the RIF patients but not healthy control, which indicates an advanced or accelerated menstrual cycle in these patients. This study provides an insight of applying the single cell technology to decipher the underlying cell composition during endometrium cycle, and directly connected the lower cellular diversity and advanced gene expression profiles to clinic symptom.

IMGT®, the international ImMunoGeneTics information system® http://www.imgt.org,is the global reference in immunogenetics and immunoinformatics, created in 1989 by Marie-Paule Lefranc (Université de Montpellier and CNRS). IMGT® is a high -quality integrated knowledge resource specialized in the immunoglobulins (IG) or antibodies , T cell receptors (TR), major histocompatibility (MH) of human and other vertebrate species , and in the immunoglobulin superfamily (IgSF), MH superfamily (MhSF) and related proteins of the immune system (RPI) of vertebrates and invertebrates. IMGT® provides a common access to sequence, genome and structure Immunogenetics data, based on the concepts of IMGT-ONTOLOGY and on the IMGT Scientific chart rules. IMGT® works in close collaboration with EBI (Europe) , DDBJ (Japan) and NCBI (USA). IMGT® consists of sequence databases, genome database, structure database, and monoclonal antibodies database, Web resources and interactive tools.

Medical research paradigm is changing, from symptom-based medicine to evidence-based medicine to precision medicine. Nowadays, at the very beginning of a medical study, the principle investigator needs to make detailed data management plan (DMP), indicating the scientific data storage, data license, data sharing, data preservation and etc. In this talk, I will introduce the efforts made by the National Scientific Data Center for Population and Health on scientific data collection, data curation, data sharing, and data management tools development. Furthermore, I’d like to discuss the technical and management challenges in data-driven medical studies.

Rapidly proliferating cancer cells have much higher demand for proteinogenic amino acids than normal cells. The use of amino acids in human proteomes is largely affected by their bioavailability, which is constrained by the biosynthetic energy cost in the living organisms. Conceptually distinct from gene-based analyses, we introduce the energy cost per amino acid (ECPA) to quantitatively characterize the use of 20 amino acids during protein synthesis in human cells. By analyzing gene expression data from The Cancer Genome Atlas, we find that cancer cells evolve to utilize amino acids more economically by optimizing gene expression profiles. We further validate this pattern in an experimental evolution of xenograft tumors. ECPA not only shows robust prognostic power across many cancer types, but also improves the prediction of tumor response to checkpoint inhibitor immunotherapy. Our ECPA analysis reveals a common principle during cancer evolution.

I will discuss our efforts in applying high-throughput DNA and RNA sequencing in the diagnostics and longitudinal monitoring of leukemias patients. We show that targeted sequencing is effective in ascertaining the mutation load after chemotherapy or allogeneic hematopoietic transplant, which in turn can serve as a good prognostic indicator. I will also discuss how such longitudinal mutation data can be used to trace the clonal evolution in leukemia patients. I will further discuss how patient’s electronic health data can be used in predicting the risk of developing leukemia.

Transposable elements can influence gene transcript ion and biological function through various mechanisms. Long terminal repeats (LTRs) of endogenous retroviruses (ERVs) have been shown to influence the expression of neighboring genes. Solitary LTRs contain various transcriptional regulatory elements including promoters, enhancers, and polyadenylation signals. They provide miRNAs during species radiation. Hypomethylation of the LTR element allows the neighboring functional gene to have tissue specific expression. Accumulated changes of the LTR elements in gene regulation are likely to be functional factors for the process of diversification, speciation and evolution consequences. A small minority of such sequences has acquired a role in regulating gene expression, and some of these may be related to differences between individuals, and to expression of disease. They seemed to be a source of alternative splicing, structural change of genomes, chromosome evolution, and could be related to genetic variation and epigenetic regulation linked to various diseases in various species.

Ribonucleic acid (RNA) molecules are master regulators of cells. They are involved in a variety of molecular processes: they transmit genetic information, sense cellular signals and communicate responses, and even catalyze chemical reactions. As in the case of proteins, RNA function is dictated by its structure and by its ability to adopt different conformations, which in turn is encoded in the sequence. Understanding how these molecules interact to carry out their biological roles requires detailed knowledge of RNA 3D structure and dynamics as well as thermodynamics, which strongly governs the folding of RNA and RNA-RNA interactions as well as a host of other interactions within the cellular environment. Experimental determination of these properties is difficult, and various computational methods have been developed to model the folding of RNA 3D structures and their interactions with other molecules. However, computational methods have limitations, especially when the biological effects demand computation of the dynamics beyond a few hundred nanoseconds. For the researcher confronted with such challenges, a more amenable approach is to resort to coarse-grained modeling to reduce the number of data points and computational demand to a more tractable size, while sacrificing as little critical information as possible.

I will review strategies for computational modeling of RNA structure and dynamics, and for the structure determination of RNA interactions and complexes with other molecules (ions, small molecules, and proteins). I will show applications of SimRNA, a method developed in my laboratory, which uses a coarse-grained representation, relies on the Monte Carlo method for sampling the conformational space, and employs a statistical potential to approximate the energy and identify conformations that correspond to biologically relevant structures. I will also discuss the use of computational approaches for RNA structure prediction that can utilize data from experimental analyses.

Large-scale outbreaks of dengue and human influenza have frequently occurred in Taiwan during summer and winter seasons, respectively. Facing challenges of fast increasing numbers of cases and deaths, an integrated epidemiological information system which provides timely data for decision-makers is necessary. Since March 2004 [i.e. after the outbreak of severe acute respiratory syndrome (SARS) in 2003], we established a hospital emergency department-based syndromic surveillance system (ED-SSS) that involves 11 syndrome groups capable of automatic transmission of daily patient data directly from hospitals with emergency care throughout Taiwan to the Centers for Disease Control in Taiwan (Taiwan-CDC). In addition, school-based infectious disease syndromic surveillance system (SID-SSS) was developed from Taipei City government in response to the 2009 influenza A (H1N1) pandemic. Teachers and nurses from preschools to universities in all 12 districts of Taipei City are required to report cases of symptomatic children or sick leaves on a daily bases through the SID-SSS. The SID-SSS has extended to New Taipei City and other parts in Taiwan. In other words, both ED-SSS and SID-SSS can provide daily epidemi-ological information when outbreaks of emerging infectious diseases caused by novel agents [e.g. enterovirus 71, 2009 swine influenza virus (pdmH1N1/09)] occur.

In 2015, the most severe and largest epidemic of dengue occurred in southern Taiwan. A risk factor-based integrated dengue information system (RF-IDIS) which involve important risk factors (e.g. mosquito indices, meteorological factors, land use, high risk locations etc.) to assist outbreak investigation and risk management. This RF-IDIS was firstly established in Pingtung in 2016 and subsequently extended to other cities/counties in Taiwan. Source reduction of Aedes mosquitoes can start early from entomological surveillance data even before the detection of human cases.

In the future, integrated influenza surveillance systems need to have human vaccine history, viral sequences of the virus derived from different host species, and ecology.

Take transmission dynamics of seasonal influenza virus as an example, I will show you the idea of using theoretical modelling to address critical questions in life and health sciences by integrating big data from multiple sources. I will show you the successes of long-term incidence forecasting of seasonal influenza virus recent years for the United States, and the challenges due to the complexity of pathogen evolution. Data integration based on mechanistic model is a way forward for study the complex diseases in life and health sciences.

With the accumulation of genotype and phenotype data over an increasing big number of individuals, genome-wide association studies have identified many genetic variants associated with complex diseases and traits. However, we lack high-throughput experimental tools to understand the biological functions of these variants. We also lack computational methods to integrate multi-omics data, or differenttypes of “big biology data”, to explore the genetic basis of complex diseases and traits. In my talk, I will present two experimental methods and one computational method that we have built/extended in the past in the field of human genetics. I will also talk a little bit about my future work focusing on utilizing big data in animal genetics to accelerate livestock genetic improvement.

Bringing together the expertise of bioinformaticians and geneticists is crucial, since very specific and fundamental computational approaches are required for virus, microorganism and human genome research, particularly in an era of big data. Improve existing analytical tools, computational resources, data sharing approaches, new diagnostic tools, and bioinformatic training are crucial. In this talk I will present results in collaboration with several research groups in Brazil related to the Zika virus, neglected and genetic diseases and resistance to antibiotics.

Background: Some researches have projected a relationship between scarlet fever and meteorology and air pollution in a specific regions or cities, but the resurge mechanism of scarlet fever after 2011 in China has not interpreted in view of nationwide.
Objectives: Aimed to investigate the association between the resurgence of scarlet fever and long-term exposure to air pollutants and weather conditions in China. Methods: Data on scarlet fever were from the National Notifiable Disease Reporting System. Air pollutants were from the National Environmental Protection Department, and weather conditions were from the National Meteorological Information Center. A lag non-linear model (DLNM) was used to estimate the excess risk of scarlet fever associated with air pollutants and weather conditions.
Results: 655,039 scarlet fever cases were reported during 2004-2018. It started to surge in 2011 (4.7638 per 100 000), and peaked in 2018 (5.6736 per 100 000). The average incidence in 2011-18 was twice that in 2004-10 (rate ratio=2.302; 95% CI: 2.289-2.314; p<0.001]. There was a low-moderate correlation between scarlet fever and monthly NO2 concentration (r=0.21), sunlight (r=0.28), and wind speed (r=0.24), but the others were weak [PM10 (r=0.13), ozone (r=0.11), and PM2.5 (r=0.05)]. By contrast, it was inversely correlated with monthly relative humidity (RH, r=-0.37), precipitation (r=-0.24) and mean temperature (r=-0.2). A one-unit increment of NO₂ concentration was associated with scarlet fever increased (RR=1.5, 95% CI: 1.35-1.66). The RR was recorded at lag 0 at 42 μg/m3 NO2 concentration (RR=1.15). NO2 pooled estimates varied substantially across China (RR=1.02~4.16), but were higher in northern parts.
Conclusions: Long-term exposure to ambient NO₂ was associated with scarlet fever resurgence, triggered by low RH, temperature, precipitation, and high wind speed and sunshine. Further interventions to reduce NO2 emissions might suppress the resurgence of scarlet fever.