Effective utilization of rapidly developing ‘omic profiling technologies and, 52in particular, the introduction of personalized/precision/preventive medi-cine have recently become major goals of medical research. This paradigm shift requires moving away from traditional approaches that do not adequately consider the individual characteristics of each patient. Our laboratory develops strategies to address these challenges by bringing the ideas and methods from mathematics and computational sciences to the medical domain. The first part of our approach is driven by integrative analysis of clinical and omic data and aims to explore the etiologies of intractable diseases. Next, we classify each disease into finer categories, such as types of anti-cancer immune responses, using molecular profiles and then we clarify underlying causal mechanisms with systems-based approaches. Lastly, we apply mathematical and machine learning techniques to infer optimal therapy for each patient to guide treatment decisions made by their hospital or clinic. Similar approaches can be used for disease prevention based on an individual’s medical history. Our past and cur-rent research projects include: (1) Investigating the relationship between tumor microenvironment, subclonal diversity, drug response, and patient prognosis in lung, colorectal and liver cancer, (2) Development and application of novel machine learning methods for cancer immunology multi-omics, (3) Integrative Trans-omics modelling of disease-associated genomic variations, (4) Accurate insertion/deletion calling from next-generation sequencing (NGS) data, (5) Whole exome sequencing (WES) analysis to identify intractable disease-causing genes, (6) Cancer Whole genome sequencing (WGS) analysis, (7) Development of new clustering methods, (8) Development of cancer classification and prog-nosis prediction methods based on gene expression data, (9) Prediction of opti-mal drug combinations for cancer chemotherapy (10) Drug toxicity prediction with machine learning, (11) Prediction of post-translational amino-acid modi-fications, protein structure, protein-peptide interactions, molecular recognition features (MoRFs), and protein functions (12) Discovery of clinically-relevant subtypes for cancer immunotherapy, and (13) Developing AI and deep learning technologies for image and ‘omic data analyses.Figure: DeepInsight – deep learning for ‘omic data analysis(a) Pipeline. Transformation from feature vector to im-age pixels. (b) Parallel convolutional neural network (CNN) architecture used in DeepInsight.Recent Major PublicationsRheinbay E, Nielsen MM, Abascal F, Wala JA, Shapira O, Tiao G, Hornshøj H, Hess JM, Juul RI, Lin Z, Feuerbach L, Sabarinathan R, Madsen T, Kim J, Mularoni L, Shuai S, Lanzós A, Herrmann C, Maruvka YE, Shen C, Amin SB, Bandopadhayay P, Bertl J, Boroevich KA, Busanovich J, Carlevaro-Fita J, Chakravarty D, Chan CWY, Craft D, Dh-ingra P, Diamanti K, Fonseca NA, Gonzalez-Perez A, Guo Q, Hamilton MP, Haradhvala NJ, Hong C, Isaev K, John-son TA, … , Tsunoda T, et al., Analyses of non-coding somatic drivers in 2,658 cancer whole genomes. Nature 578, 102-111 (2020)Choobdar S, Ahsen ME, Crawford J, Tomasoni M, Fang T, Lamparter D, et al., Assessment of network module identification across complex diseases. Nature Methods 16, 843-852 (2019)Sharma A*, Vans E, Shigemizu D, Boroevich KA, Tsunoda T*. DeepInsight: a methodology to transform a non-image data to an image for convolution neural network architecture. Scientific Reports 9, 11399 (2019)Invited presentationsSharma A. “CNN to non-image data by DeepInsight method” MATLAB EXPO (Japan/Online) September 2020Tsunoda T. “Personalized Cancer Medicine with Het-erogeneity and Immunological Analysis.” CREST Inter-national Symposium on Big Data Application (Tokyo, Japan) January 2020Laboratory for Medical Science MathematicsTeam Leader: Tatsuhiko Tsunoda
元のページ ../index.html#58