Year: 2021
Type: dissertation
Abstract: In modern biomedicine, the role of computation becomes more crucial in light of the ever-increasing growth of biological data, which requires effective computational methods to integrate them in a meaningful way and unveil previously undiscovered biological insights.In this dissertation, we introduce a series of machine learning algorithms for biomedical discovery.Focused on protein functions in the context of system biology, these machine learning algorithms learn representations of protein sequences, structures, and networks in both the small-and large-data scenarios.First, we present a deep learning model that learns evolutionary contexts integrated representations of protein sequence and assists to discover protein variants with enhanced functions in protein engineering.Second, we describe a geometric deep learning model that learns representations of protein and compound structures to inform the prediction of protein-compound binding affinity.Third, we introduce a machine learning algorithm to integrate heterogeneous networks by learning compact network representations and to achieve drug repurposing by predicting novel drug-target interaction.We also present new scientific discoveries enabled by these machine learning algorithms.Taken together, this dissertation demonstrates the potential of machine learning to address the small-and large-data challenges of biomedical data and transform data into actionable insights and new discoveries.During my Ph.D. study, I have been enjoying reading the acknowledgments section in others' dissertations.There were supportive people and vivid stories behind everyone's journey of a Ph.D., and mine was not an exception.I have been extremely lucky to be surrounded by a number of advisors, colleagues, friends, and family both in and out of the University of Illinois Urbana-Champaign (UIUC).And now, I finally have the opportunity to write down my acknowledgments section to thank them all, with my greatest gratitude.First and foremost, I would like to give my warmest thanks to my advisor, Professor Jian Peng.I still remember the first meeting I had with Jian during his visit to Tsinghua University in 2015.He told me, with enthusiasm and excitement, about his new algorithm for identifying target genes of Parkinson's disease.I was immediately impressed by his research and the potential impact that could be achieved.That summer, Jian hosted me as a visiting undergraduate student at UIUC, and I had a fun time working with him on two research projects, part of which was then further developed into some results presented in this dissertation.The experience was so joyous that it became a no-brainer for me to choose Jian as my advisor when I was applying to graduate schools.Throughout my five years at UIUC, Jian has been a never-ending source of guidance, support, and inspiration for me and kept passing his knowledge, caring, and encouragement to me.He is not only a visionary advisor with insightful advice and sharp questions but also a great colleague who knows the finest details of a problem and could help with any technical challenges I had encountered.Now when I am reading the thousands of emails and Skype text messages I exchanged with him over the past few years, plenty of great memories flashback: the detailed notes he sent about the trip from Chicago airport to Champaign before my first visit to UIUC, the extensive discussion when we were working on our first RECOMB submission, the numerous paper links he sent to guide me to get familiar with a new research topic, the clever ideas he proposed and key issues he pinpointed when I was struggling with research, his affirmative "sure, always happy to" every time I requested help from him, and his huge effort in supporting my job search.Those good old times very well illustrate how I become who I am today after five years working with Jian.During my Ph.D., Jian gave me tremendous freedom, which was truly a luxury, and I greatly appreciate his patience and belief when helping me grow my research independence.Looking back, what I learned from Jian is far beyond a degree, but the way to think, the philosophy to do research, and the career to pursue as a scholar.It was my privilege to have Jian as my advisor, and I am forever iii grateful to him.I also owe many thanks to my other mentors who have guided and led me through my Ph.D. I have special gratitude for Professor Jianzhu Ma.I started to work with Jianzhu in my second year and he remains my close collaborator since then.While we only met once in person during my Ph.D., we made innumerable Skype calls to brainstorm research ideas and discuss coding and writing.Working with Jianzhu was a precious experience for me and it really motivated me to think about how to formulate research problems.Thanks also go to Professor Mohammed El-Kebir.I worked with Mohammed on cancer genomics for a period of time.He is an amazing mentor on everything, from research, to writing and presentation, and to coding and paper reviewing.I will always remember the scenes he welcomed me into his office "Yunan, come on in!" and started a fruitful meeting with me.Working with Professor Kaiyu Guan on remote sensing and intelligent agriculture was another unique experience.I was strongly impressed and encouraged by his scholarship: self-discipline, determination, and energy.I believe that everyone knowing Kaiyu, including me, might never understand how one can accomplish so many things at the same time.I also closely worked with Professor Huimin Zhao on protein engineering, who taught me how to study scientific problems in an interdisciplinary context.It was my pleasure to collaborate with Professor Bonnie Berger in early years of my Ph.D., and I benefited a lot from every stage of the project.I would also like to thank many other brilliant professors and researchers.I want to especially thank Professor Jianyang Zeng for introducing me to the field of computational biology when I was a sophomore at Tsinghua University and for providing me with great support even after I started my Ph.D. in the US.Special thanks also go to Professor Andrew Chi-Chih Yao for founding Yao Class and for creating opportunities for undergraduate students like me to do frontier research.It is my honor to have Professor Jiawei Han on my dissertation committee, and his diligence and passion for research have a deep influence on me.Thanks to Dr. Hoon Cho for being my external committee member.His paper with Jian was the first RECOMB paper I have read, which also inspired part of the work in this dissertation.I am also grateful to Professors Tandy Warnow and Saurabh Sinha for providing me with invaluable advice at many stages of my degree.Dr. Ali Madani hosted me remotely at Salesforce Research for a summer internship, which was truly a wonderful experience.During my job search, I was fortunate to have met with and received guidance from many researchers both at UIUC and other institutions
Author: Luo, Yunan
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7057918551