DOCSLIB.ORG

2014 ISCB Accomplishment by a Senior Scientist Award: Gene Myers

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
2014 ISCB Accomplishment by a Senior Scientist Award: Gene Myers Message from ISCB 2014 ISCB Accomplishment by a Senior Scientist Award: Gene Myers Christiana N. Fogg1, Diane E. Kovats2* 1 Freelance Science Writer, Kensington, Maryland, United States of America, 2 Executive Director, International Society for Computational Biology, La Jolla, California, United States of America The International Society for Computa- tional Biology (ISCB; http://www.iscb. org) annually recognizes a senior scientist for his or her outstanding achievements. The ISCB Accomplishment by a Senior Scientist Award honors a leader in the field of computational biology for his or her significant contributions to the com- munity through research, service, and education. Dr. Eugene ‘‘Gene’’ Myers of the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden has been selected as the 2014 ISCB Accom- plishment by a Senior Scientist Award winner. Myers (Image 1) was selected by the ISCB’s awards committee, which is chaired by Dr. Bonnie Berger of the Massachusetts Institute of Technology (MIT). Myers will receive his award and deliver a keynote address at ISCB’s 22nd Image 1. Gene Myers. Image credit: Matt Staley, HHMI. Annual Intelligent Systems for Molecular doi:10.1371/journal.pcbi.1003621.g001 Biology (ISMB) meeting. This meeting is being held in Boston, Massachusetts, on July 11–15, 2014, at the John B. Hynes guidance of his dissertation advisor, An- sequences and how to build evolutionary Memorial Convention Center (https:// drzej Ehrenfeucht, who had eclectic inter- trees. www.iscb.org/ismb2014). ests that included molecular biology. Myers landed his first faculty position in Myers was captivated by computer Myers, along with fellow graduate students the Department of Computer Science at programming as a young student. He and future bioinformaticians Gary Stormo the University of Arizona. Throughout his remembered his fascination being stoked and David Haussler, was drawn by research career, he has been interested in by a realization and recalls, ‘‘A computer Ehrenfeucht’s curiosity about such basic sequence assembly. ‘‘I first started thinking is a programmable device, and once questions as how to compare DNA about this problem in 1984,’’ he recollected. programmed, is a specific device for doing something that I conceived of. I found this magical.’’ He completed a BS in mathe- Citation: Fogg CN, Kovats DE (2014) 2014 ISCB Accomplishment by a Senior Scientist Award: Gene Myers. PLoS matics at the California Institute of Comput Biol 10(5): e1003621. doi:10.1371/journal.pcbi.1003621 Technology, but his interest in biology Published May 22, 2014 came during his PhD studies at the Copyright: ß 2014 Fogg, Kovats. This is an open-access article distributed under the terms of the Creative University of Colorado in the late 1970s. Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Myers recounted that he initially consid- ered molecular biology as ‘‘a source of Funding: CNF was paid by ISCB to write this article. interesting computational questions.’’ He Competing Interests: The authors have declared that no competing interests exist. was studying computer science under the * E-mail: [email protected] PLOS Computational Biology | www.ploscompbiol.org 1 May 2014 | Volume 10 | Issue 5 | e1003621 ‘‘While I developed many seminal algo- their function to produce tissues of a much, much faster way to get it done.’’ rithms for sequence comparison and given function and shape. Genomes Despite the pros and cons associated search in the ’80s and early ’90s, encode this; how?’’ with both of these settings, Myers feels including BLAST (Basic Local Alignment Myers acknowledged that microscopy that ‘‘private universities or not-for-profit Search Tool), the problem that has and and computational bioimaging are hin- institutions, such as the Max-Planck- continues to fascinate me to this day is dered by limits of image resolution, the Gesselschaft (MPG) that I belong to now, sequence assembly.’’ He is well-known often weak signals produced by imaging are terrific in that there is more flexi- for being one of the authors of the 1990 reagents, and the inability to robotically bility of funding, and thus, it is easier to manuscript that first described BLAST, a control commercial microscopes. Myers take risks and work towards long-term groundbreaking algorithm that is still and his research group are working to visions.’’ used today for sequence comparison. overcome some of these barriers by Myers’s unique contributions to com- This paper is also one of the most cited building their own microscopic devices. putational biology have been recognized papers in scientific literature. He sees this type of work as having the by several awards, including election to the Myers’s deep-seated interest in se- potential to revolutionize medicine. He National Academy of Engineering (2003), quence assembly led him to promote the said, ‘‘Really understanding (in molecular the Association for Computing Machinery idea that whole genome shotgun sequenc- terms) what a cell can do and how what it (ACM) Kannellakis Prize (2002), and the ing, which had been effective in sequenc- does affects its role in a complex tissue or International Max Planck Research Prize ing microbial genomes, could be used on organ will greatly advance medicine and (2004). Bonnie Berger (MIT), chair of the the large and unwieldy human genome. treatment as well as help us understand ISCB Nominating Committee, sees Myers Craig Venter believed in this approach as variation across species and how organ- as an exemplar of the Achievement by a well, and he brought Myers to Celera isms develop.’’ Senior Scientist Award. Berger stated that, Genomics in 1998 during their push to Myers recounted the importance of key ‘‘Myers is one of the founders of the field, sequence the human genome. Myers mentors in the success of his career. Myers bringing his algorithmic expertise to the recalled writing thousands of lines of code met Webb Miller when he was a young most fundamental problems in computa- to build algorithms that could assemble faculty member at the University of tional biology. From his role in creating the vast amounts of sequence data. He Arizona in the early 1980s. The two struck the indispensable and widely used BLAST considers the success of this landmark up a fruitful collaboration that led to many program for basic sequence search, to sequencing project as a highlight of his early papers about sequence analysis. breaking the barrier of sequencing the career. Myers explained, ‘‘Miller helped me human genome, to deciphering what is In 2002, Myers returned to academia in greatly in the early part of my career in coded in DNA, he has launched our a position at the University of California, that he taught me, through example, that discipline. Myers has been a prominent Berkeley’s Center for Integrative Geno- writing can be fun.’’ Myers gained a member of the ISCB community, serving mics. Several animal genomes had been different sort of insight while working for on the Board of Directors, as an ISCB sequenced, and many more were in Venter at Celera. He described Venter as fellow, and as chair and area chair for progress. Myers and others were getting ‘‘a master of the sound bite, and while this numerous ISMB meetings.’’ Alfonso Va- to work on mining the mounting collection may sound trivial, it is actually more lencia, leader of the Structural Computa- of genetic information and developing important than one might think. Much tional Biology group at the Spanish algorithms to compare genomes at an of one’s career success depends on the National Cancer Research Center and unprecedented scale. ability to present one’s ideas in powerful, president-elect of ISCB, also sees Myers as More recently, Myers headed a lab at succinct, clear ways.’’ a stellar representative of the field. Valen- the Howard Hughes Medical Institute Myers believes that mentorship should cia said of this year’s award winner, ‘‘I am (HHMI)’s Janelia Farm Research Cam- be ‘‘about shaping the character of the particularly happy about the election of pus, and in 2012, he moved to Dresden, individual and their understanding of their Gene Myers, since he represents the strong Germany, and became director at the role within the research community.’’ He roots of computational biology in algorith- Max Planck Institute of Molecular Cell starts with himself as he aims to ‘‘do [his] mic and method development. The inten- Biology and Genetics and the Klaus- best to be a good role model, to instill sity with which he lives science, the Tschira chair of the Systems Biology values of integrity, objectivity, and open- originality of his approaches, and the Center. Myers describes that his ‘‘latest ness.’’ He states, ‘‘I see my role as mentor attention he dedicates to the technical focus all started because [he] wish[es], like as simply to create a vibrant intellectual details are characteristics of his work and a many, to ‘decode’ the genome.’’ His work atmosphere that allows my students to great example for our new generations of has evolved into building microscopic blossom.’’ Myers has trained students from bioinformaticians and computational biol- devices and image analysis tools that can varied academic backgrounds throughout ogists.’’ be used to observe and model the inner his career, and he considers that passion as Myers remains fervent and passionate workings of cells and biological systems. well as talent are vital to the success of his about the work he does. He contends that ‘‘Having the genome of an organism trainees. Myers affirmed that ‘‘there is no his upbringing, which included traveling allows one to engage in transgenics on a substitute for passion.’’ the globe with his family, as well as his systematic, genome-wide scale,’’ he ex- Myers has also benefitted by working innate passion for science and mathemat- plained.
Load more

Recommended publications
  • Algorithms for Computational Biology 8Th International Conference, Alcob 2021 Missoula, MT, USA, June 7–11, 2021 Proceedings
    Lecture Notes in Bioinformatics 12715 Subseries of Lecture Notes in Computer Science Series Editors Sorin Istrail Brown University, Providence, RI, USA Pavel Pevzner University of California, San Diego, CA, USA Michael Waterman University of Southern California, Los Angeles, CA, USA Editorial Board Members Søren Brunak Technical University of Denmark, Kongens Lyngby, Denmark Mikhail S. Gelfand IITP, Research and Training Center on Bioinformatics, Moscow, Russia Thomas Lengauer Max Planck Institute for Informatics, Saarbrücken, Germany Satoru Miyano University of Tokyo, Tokyo, Japan Eugene Myers Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany Marie-France Sagot Université Lyon 1, Villeurbanne, France David Sankoff University of Ottawa, Ottawa, Canada Ron Shamir Tel Aviv University, Ramat Aviv, Tel Aviv, Israel Terry Speed Walter and Eliza Hall Institute of Medical Research, Melbourne, VIC, Australia Martin Vingron Max Planck Institute for Molecular Genetics, Berlin, Germany W. Eric Wong University of Texas at Dallas, Richardson, TX, USA More information about this subseries at http://www.springer.com/series/5381 Carlos Martín-Vide • Miguel A. Vega-Rodríguez • Travis Wheeler (Eds.) Algorithms for Computational Biology 8th International Conference, AlCoB 2021 Missoula, MT, USA, June 7–11, 2021 Proceedings 123 Editors Carlos Martín-Vide Miguel A. Vega-Rodríguez Rovira i Virgili University University of Extremadura Tarragona, Spain Cáceres, Spain Travis Wheeler University of Montana Missoula, MT, USA ISSN 0302-9743 ISSN 1611-3349 (electronic) Lecture Notes in Bioinformatics ISBN 978-3-030-74431-1 ISBN 978-3-030-74432-8 (eBook) https://doi.org/10.1007/978-3-030-74432-8 LNCS Sublibrary: SL8 – Bioinformatics © Springer Nature Switzerland AG 2021 This work is subject to copyright.
    [Show full text]
  • Computational Pan-Genomics: Status, Promises and Challenges
    bioRxiv preprint doi: https://doi.org/10.1101/043430; this version posted March 12, 2016. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Computational Pan-Genomics: Status, Promises and Challenges Tobias Marschall1,2, Manja Marz3,60,61,62, Thomas Abeel49, Louis Dijkstra6,7, Bas E. Dutilh8,9,10, Ali Ghaffaari1,2, Paul Kersey11, Wigard P. Kloosterman12, Veli M¨akinen13, Adam Novak15, Benedict Paten15, David Porubsky16, Eric Rivals17,63, Can Alkan18, Jasmijn Baaijens5, Paul I. W. De Bakker12, Valentina Boeva19,64,65,66, Francesca Chiaromonte20, Rayan Chikhi21, Francesca D. Ciccarelli22, Robin Cijvat23, Erwin Datema24,25,26, Cornelia M. Van Duijn27, Evan E. Eichler28, Corinna Ernst29, Eleazar Eskin30,31, Erik Garrison32, Mohammed El-Kebir5,33,34, Gunnar W. Klau5, Jan O. Korbel11,35, Eric-Wubbo Lameijer36, Benjamin Langmead37, Marcel Martin59, Paul Medvedev38,39,40, John C. Mu41, Pieter Neerincx36, Klaasjan Ouwens42,67, Pierre Peterlongo43, Nadia Pisanti44,45, Sven Rahmann29, Ben Raphael46,47, Knut Reinert48, Dick de Ridder50, Jeroen de Ridder49, Matthias Schlesner51, Ole Schulz-Trieglaff52, Ashley Sanders53, Siavash Sheikhizadeh50, Carl Shneider54, Sandra Smit50, Daniel Valenzuela13, Jiayin Wang70,71,72, Lodewyk Wessels56, Ying Zhang23,5, Victor Guryev16,12, Fabio Vandin57,34, Kai Ye68,69,72 and Alexander Sch¨onhuth5 1Center for Bioinformatics, Saarland University, Saarbr¨ucken, Germany; 2Max Planck Institute for Informatics, Saarbr¨ucken,
    [Show full text]
  • Department of Energy Office of Health and Environmental Research SEQUENCING the HUMAN GENOME Summary Report of the Santa Fe Workshop March 3-4, 1986
    Department of Energy Office of Health and Environmental Research SEQUENCING THE HUMAN GENOME Summary Report of the Santa Fe Workshop March 3-4, 1986 Los Alamos National Laboratory Los Alamos Los Alamos, New Mexico 87545 Los Alamos National Laboratory is operated by the University of California for the United States Department of Energy under contract W-7405-ENG-36. DEPARTMENT OF ENERGY OFFICE OF HEALTH AND ENVIRONMENTAL RESEARCH SEQUENCING THE HUMAN GENOME SUMMARY REPORT ON THE SANTA FE WORKSHOP (MARCH 3-4, 1986) Executive Summary. The following is a summary of the Santa Fe Workshop held on March 3 and 4, 1986. The workshop was sponsored by the Office of Health and Environmental Research (OHER) and Los Alamos National Laboratory (LANL) and dedicated to examining the feasibility, advisability, and approaches to sequencing the human genome. The workshop considered four principal topics: I. Technologies to be employed. II. Expected benefits. III. Architecture of the enterprise. IV. Participants and funding. I . Technology The participants of the workshop foresaw extraordinary and continuing progress in the efficiency and accuracy of mapping, ordering , and sequencing technologies. They suggested that a coordinated analysis of the human genome begin with the task of ordering overlapping recombinant DNA fragments obtained from purified human chromosomes that would provide an infrastructure for sequencing activity. At the same time, they support in-depth evaluation of current and developing strategies for sequencing including possible applications of automation and robotics that would minimize the time and cost of sequencing. II. Benefits The socio-political and health benefits, and the benefit:cost ratio were seen as highly favorable not only for human health, but in addition for the development of new diagnostic, preventative and therapeutic tools, jobs, and industries.
    [Show full text]
  • 120421-24Recombschedule FINAL.Xlsx
    Friday 20 April 18:00 20:00 REGISTRATION OPENS in Fira Palace 20:00 21:30 WELCOME RECEPTION in CaixaForum (access map) Saturday 21 April 8:00 8:50 REGISTRATION 8:50 9:00 Opening Remarks (Roderic GUIGÓ and Benny CHOR) Session 1. Chair: Roderic GUIGÓ (CRG, Barcelona ES) 9:00 10:00 Richard DURBIN The Wellcome Trust Sanger Institute, Hinxton UK "Computational analysis of population genome sequencing data" 10:00 10:20 44 Yaw-Ling Lin, Charles Ward and Steven Skiena Synthetic Sequence Design for Signal Location Search 10:20 10:40 62 Kai Song, Jie Ren, Zhiyuan Zhai, Xuemei Liu, Minghua Deng and Fengzhu Sun Alignment-Free Sequence Comparison Based on Next Generation Sequencing Reads 10:40 11:00 178 Yang Li, Hong-Mei Li, Paul Burns, Mark Borodovsky, Gene Robinson and Jian Ma TrueSight: Self-training Algorithm for Splice Junction Detection using RNA-seq 11:00 11:30 coffee break Session 2. Chair: Bonnie BERGER (MIT, Cambrige US) 11:30 11:50 139 Son Pham, Dmitry Antipov, Alexander Sirotkin, Glenn Tesler, Pavel Pevzner and Max Alekseyev PATH-SETS: A Novel Approach for Comprehensive Utilization of Mate-Pairs in Genome Assembly 11:50 12:10 171 Yan Huang, Yin Hu and Jinze Liu A Robust Method for Transcript Quantification with RNA-seq Data 12:10 12:30 120 Zhanyong Wang, Farhad Hormozdiari, Wen-Yun Yang, Eran Halperin and Eleazar Eskin CNVeM: Copy Number Variation detection Using Uncertainty of Read Mapping 12:30 12:50 205 Dmitri Pervouchine Evidence for widespread association of mammalian splicing and conserved long range RNA structures 12:50 13:10 169 Melissa Gymrek, David Golan, Saharon Rosset and Yaniv Erlich lobSTR: A Novel Pipeline for Short Tandem Repeats Profiling in Personal Genomes 13:10 13:30 217 Rory Stark Differential oestrogen receptor binding is associated with clinical outcome in breast cancer 13:30 15:00 lunch break Session 3.
    [Show full text]
  • BIOGRAPHICAL SKETCH NAME: Berger
    BIOGRAPHICAL SKETCH NAME: Berger, Bonnie eRA COMMONS USER NAME (credential, e.g., agency login): BABERGER POSITION TITLE: Simons Professor of Mathematics and Professor of Electrical Engineering and Computer Science EDUCATION/TRAINING (Begin with baccalaureate or other initial professional education, such as nursing, include postdoctoral training and residency training if applicable. Add/delete rows as necessary.) EDUCATION/TRAINING DEGREE Completion (if Date FIELD OF STUDY INSTITUTION AND LOCATION applicable) MM/YYYY Brandeis University, Waltham, MA AB 06/1983 Computer Science Massachusetts Institute of Technology SM 01/1986 Computer Science Massachusetts Institute of Technology Ph.D. 06/1990 Computer Science Massachusetts Institute of Technology Postdoc 06/1992 Applied Mathematics A. Personal Statement Advances in modern biology revolve around automated data collection and sharing of the large resulting datasets. I am considered a pioneer in the area of bringing computer algorithms to the study of biological data, and a founder in this community that I have witnessed grow so profoundly over the last 26 years. I have made major contributions to many areas of computational biology and biomedicine, largely, though not exclusively through algorithmic innovations, as demonstrated by nearly twenty thousand citations to my scientific papers and widely-used software. In recognition of my success, I have just been elected to the National Academy of Sciences and in 2019 received the ISCB Senior Scientist Award, the pinnacle award in computational biology. My research group works on diverse challenges, including Computational Genomics, High-throughput Technology Analysis and Design, Biological Networks, Structural Bioinformatics, Population Genetics and Biomedical Privacy. I spearheaded research on analyzing large and complex biological data sets through topological and machine learning approaches; e.g.
    [Show full text]
  • Mapping Our Genes—Genome Projects: How Big? How Fast?
    Mapping Our Genes—Genome Projects: How Big? How Fast? April 1988 NTIS order #PB88-212402 Recommended Citation: U.S. Congress, Office of Technology Assessment, Mapping Our Genes-The Genmne Projects.’ How Big, How Fast? OTA-BA-373 (Washington, DC: U.S. Government Printing Office, April 1988). Library of Congress Catalog Card Number 87-619898 For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402-9325 (order form can be found in the back of this report) Foreword For the past 2 years, scientific and technical journals in biology and medicine have extensively covered a debate about whether and how to determine the function and order of human genes on human chromosomes and when to determine the sequence of molecular building blocks that comprise DNA in those chromosomes. In 1987, these issues rose to become part of the public agenda. The debate involves science, technol- ogy, and politics. Congress is responsible for ‘(writing the rules” of what various Federal agencies do and for funding their work. This report surveys the points made so far in the debate, focusing on those that most directly influence the policy options facing the U.S. Congress, The House Committee on Energy and Commerce requested that OTA undertake the project. The House Committee on Science, Space, and Technology, the Senate Com- mittee on Labor and Human Resources, and the Senate Committee on Energy and Natu- ral Resources also asked OTA to address specific points of concern to them. Congres- sional interest focused on several issues: ● how to assess the rationales for conducting human genome projects, ● how to fund human genome projects (at what level and through which mech- anisms), ● how to coordinate the scientific and technical programs of the several Federal agencies and private interests already supporting various genome projects, and ● how to strike a balance regarding the impact of genome projects on international scientific cooperation and international economic competition in biotechnology.
    [Show full text]
  • Characterizing the Dna-Binding Site Specificities of Cis2his2 Zinc Fingers
    MQP-ID-DH-UM1 C H A R A C T E RI Z IN G T H E DN A-BINDIN G SI T E SPE C I F I C I T I ES O F C IS2H IS2 Z IN C F IN G E RS A Major Qualifying Project Report Submitted to the Faculty of the WORCESTER POLYTECHNIC INSTITUTE in partial fulfillment of the requirements for the Degrees of Bachelor of Science in Biochemistry and Biology and Biotechnology by _________________________ Heather Bell April 26, 2012 APPROVED: ____________________ ____________________ ____________________ Scot Wolfe, PhD Destin Heilman, PhD David Adams, PhD Gene Function and Exp. Biochemistry Biology and Biotech UMass Medical School WPI Project Advisor WPI Project Advisor MAJOR ADVISOR A BST R A C T The ability to modularly assemble Zinc Finger Proteins (ZFPs) as well as the wide variety of DNA sequences they can recognize, make ZFPs an ideal framework to design novel DNA-binding proteins. However, due to the complexity of the interactions between residues in the ZF recognition helix and the DNA-binding site there is currently no comprehensive recognition code that would allow for the accurate prediction of the DNA ZFP binding motifs or the design of novel ZFPs for a desired target site. Through the analysis of the DNA-binding site specificities of 98 ZFP clones, determined through a bacterial one-hybrid selection system, a predictive model was created that can accurately predict the binding site motifs of novel ZFPs. 2 T A B L E O F C O N T E N TS Signature Page ««««««««««««««««««««««««««« $EVWUDFW«««««««««««««««««««««««««««««« 7DEOHRI&RQWHQWV«««««««««««««««««««««««««« $FNQRZOHGJHPHQWV««««««««««««««««««««««««« %DFNJURXQG«««««««««««««««««««««««««««« Project Purpose «««««««««««««««««««««««««««15 0HWKRGV««««««««««««««««««««««««««««««16 5HVXOWV««««««««««««««««««««««««««««««21 'LVFXVVLRQ«««««««««««««««««««««««««««««28 Bibliograph\«««««««««««««««««««««««««««« 6XSSOHPHQWDO««««««««««««««««««««««««««« 3 A C K N O W L E D G E M E N TS I would like to thank Dr.
    [Show full text]
  • Big Data, Moocs, and ... (PDF)
    HHMI Constellation Studios for Science Education November 13-15, 2015 | HHMI Headquarters | Chevy Chase, MD Big Data, MOOCs, and Quantitative Education for Biologists Co-Chairs Pavel Pevzner, University of California- San Diego Sarah Elgin, Washington University Studio Objectives Discuss existing challenges in bioinformatics education with experts in computational biology and quantitative biology education, Evaluate best practices in teaching quantitative and computational biology, and Collaborate with scientist educators to develop instructional modules to support a biology curriculum that includes quantitative approaches. Friday | November 13 4:00 pm Arrival Registration Desk 5:30 – 6:00 pm Reception Great Hall 6:00 – 7:00 pm Dinner Dining Room 7:00 – 7:15 pm Welcome K202 David Asai, HHMI Cynthia Bauerle, HHMI Pavel Pevzner, University of California-San Diego Sarah Elgin, Washington University Alex Hartemink, Duke University 7:15 – 8:00 pm How to Maximize Interaction and Feedback During the Studio K202 Cynthia Bauerle and Sarah Simmons, HHMI 8:00 – 9:00 pm Keynote Presentation K202 "Computing + Biology = Discovery" Speakers: Ran Libeskind-Hadas, Harvey Mudd College Eliot Bush, Harvey Mudd College 9:00 – 11:00 pm Social The Pilot Saturday | November 14 7:30 – 8:15 am Breakfast Dining Room 8:30 – 10:00 am Lecture session 1 K202 Moderator: Pavel Pevzner 834a-854a “How is body fat regulated?” Laurie Heyer, Davidson College 856a-916a “How can we find mutations that cause cancer?” Ben Raphael, Brown University “How does a tumor evolve over time?” 918a-938a Russell Schwartz, Carnegie Mellon University “How fast do ribosomes move?” 940a-1000a Carl Kingsford, Carnegie Mellon University 10:05 – 10:55 am Breakout working groups Rooms: S221, (coffee available in each room) N238, N241, N140 1.
    [Show full text]
  • UC Irvine UC Irvine Previously Published Works
    UC Irvine UC Irvine Previously Published Works Title The capacity of feedforward neural networks. Permalink https://escholarship.org/uc/item/29h5t0hf Authors Baldi, Pierre Vershynin, Roman Publication Date 2019-08-01 DOI 10.1016/j.neunet.2019.04.009 License https://creativecommons.org/licenses/by/4.0/ 4.0 Peer reviewed eScholarship.org Powered by the California Digital Library University of California THE CAPACITY OF FEEDFORWARD NEURAL NETWORKS PIERRE BALDI AND ROMAN VERSHYNIN Abstract. A long standing open problem in the theory of neural networks is the devel- opment of quantitative methods to estimate and compare the capabilities of different ar- chitectures. Here we define the capacity of an architecture by the binary logarithm of the number of functions it can compute, as the synaptic weights are varied. The capacity provides an upperbound on the number of bits that can be extracted from the training data and stored in the architecture during learning. We study the capacity of layered, fully-connected, architectures of linear threshold neurons with L layers of size n1, n2,...,nL and show that in essence the capacity is given by a cubic polynomial in the layer sizes: L−1 C(n1,...,nL) = Pk=1 min(n1,...,nk)nknk+1, where layers that are smaller than all pre- vious layers act as bottlenecks. In proving the main result, we also develop new techniques (multiplexing, enrichment, and stacking) as well as new bounds on the capacity of finite sets. We use the main result to identify architectures with maximal or minimal capacity under a number of natural constraints.
    [Show full text]
  • Computational Biology and Bioinformatics
    Vol. 30 ISMB 2014, pages i1–i2 BIOINFORMATICS EDITORIAL doi:10.1093/bioinformatics/btu304 Editorial This special issue of Bioinformatics serves as the proceedings of The conference used a two-tier review system, a continuation the 22nd annual meeting of Intelligent Systems for Molecular and refinement of a process begun with ISMB 2013 in an effort Biology (ISMB), which took place in Boston, MA, July 11–15, to better ensure thorough and fair reviewing. Under the revised 2014 (http://www.iscb.org/ismbeccb2014). The official confer- process, each of the 191 submissions was first reviewed by at least ence of the International Society for Computational Biology three expert referees, with a subset receiving between four and (http://www.iscb.org/), ISMB, was accompanied by 12 Special eight reviews, as needed. These formal reviews were frequently Interest Group meetings of one or two days each, two satellite supplemented by online discussion among reviewers and Area meetings, a High School Teachers Workshop and two half-day Chairs to resolve points of dispute and reach a consensus on tutorials. Since its inception, ISMB has grown to be the largest each paper. Among the 191 submissions, 29 were conditionally international conference in computational biology and bioinfor- accepted for publication directly from the first round review Downloaded from matics. It is expected to be the premiere forum in the field for based on an assessment of the reviewers that the paper was presenting new research results, disseminating methods and tech- clearly above par for the conference. A subset of 16 papers niques and facilitating discussions among leading researchers, were viewed as potentially in the top tier but raised significant practitioners and students in the field.
    [Show full text]
  • An Efficient, Scalable and Exact Representation of High-Dimensional Color Information Enabled Via De Bruijn Graph Search
    bioRxiv preprint doi: https://doi.org/10.1101/464222; this version posted November 7, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. An Efficient, Scalable and Exact Representation of High-Dimensional Color Information Enabled via de Bruijn Graph Search Fatemeh Almodaresi1, Prashant Pandey1, Michael Ferdman1, Rob Johnson2;1, and Rob Patro1 1 Computer Science Dept., Stony Brook University ffalmodaresit,ppandey,mferdman,[email protected] 2 VMware Research [email protected] Abstract. a The colored de Bruijn graph (cdbg) and its variants have become an important combinatorial structure used in numerous areas in genomics, such as population-level variation detection in metagenomic samples, large scale sequence search, and cdbg-based reference sequence indices. As samples or genomes are added to the cdbg, the color information comes to dominate the space required to represent this data structure. In this paper, we show how to represent the color information efficiently by adopting a hierarchical encoding that exploits correlations among color classes | patterns of color occurrence | present in the de Bruijn graph (dbg). A major challenge in deriving an efficient encoding of the color information that takes advantage of such correlations is determining which color classes are close to each other in the high-dimensional space of possible color patterns. We demonstrate that the dbg itself can be used as an efficient mechanism to search for approximate nearest neighbors in this space.
    [Show full text]
  • ISMB 99 August 6 – 10, 1999 Heidelberg, Germany the Seventh
    ______________________________________ Welcome to ISMB 99 August 6 – 10, 1999 Heidelberg, Germany The Seventh International Conference on Intelligent Systems for Molecular Biology ______________________________________ Final Program and Detailed Schedule Friday, August 6, 1999 Tutorial Day The tutorials will take place in the following rooms: 8:30 – 12:30 (Coffee break around 10:30) Tutorial #1 Trübnersaal Piere Baldi Probabilistic graphical models Tutorial #2 Robert-Schumann-Zimmer Douglas L. Brutlag Bioinformatics and Molecular Biology Tutorial #3 Ballsaal Martin Reese The challenge of annotating a complete eukaryotic genome: A case study in Drosophila melanogaster Tutorial #4 Gustav-Mahler-Zimmer Tandy Warnow Computational and statistical Junhyong Kim challenges involved in reconstructing evolutionary trees Tutorial #5 Sebastian-Münster-Saal Thomas Werner The biology and bioinformatics of regulatory regions in genomes Lunch (on this day served in "Grosser Saal" on the ground floor) 13:30 – 17:30 (Coffee break around 15:30) Tutorial #6 Sebastian-Münster-Saal Rob Miller EST Clustering Alan Christoffels Winston Hide Tutorial #7 Trübnersaal Kevin Karplus Getting the most out of hidden Markov Melissa Cline models Christian Barrett Tutorial #8 Robert-Schumann-Zimmer Arthur Lesk Sequence-structure relationships and evolutionary structure changes in proteins Tutorial #9 Gustav-Mahler-Zimmer David States PERL abstractions for databases and Brian Dunford distributed computing Shore Tutorial # 10 Ballsaal Zoltan Szallasi Genetic network analysis
    [Show full text]