Introduction to Computational Biology

Tentative Schedule:

Course Description

In this course, students will develop an understanding of the key computational challenges in molecular biology. We will discuss algorithms used to solve some of these problems, and we will introduce ongoing areas of research in the fields of bioinformatics and computational biology. Grading will be based on homework assignments, two in-class quizzes, a written course paper/project, and a project presentation. Students will also be expected to contribute to class discussion and to read supplementary background materials as they find necessary.

Course Requirements:

• Prerequisites: Comp15 and at least one 100-level computer science class, or permission of the instructor. Graduate standing in computer science or a related field may be sufficient with no further prerequisites; check with the instructor. Comfort writing programs in some language is essential, as we will be implementing some of the algorithms we discuss. In the past, students have successfully used Perl, Python, Java, C, and C++. If you have another preference, please discuss your choice of language with the TA.

• Readings: The course textbook is Bioinformatics: Sequence and Genome Analysis, Second Edition by David Mount, published by Cold Spring Harbor Press. Softcover versions are available in the Medford campus bookstore. If you'd rather have the hardcover edition (expensive but more durable), you can order it online. Note that the First Edition of the book has different page numbers for the readings and does not cover some topics (such as microarrays) at all.

  The textbook was chosen because it covers, in some detail, most of the material we will discuss in the class. Readings from this text will be listed in the syllabus where appropriate. However, for individual topics, other textbooks may be better. Thus, supplementary readings from some of the recommended textbooks listed below will be listed as well. If you have no biology background, you may want to supplement the readings as well by getting a good introductory molecular biology text. (Several are available online via PubMed/Books; we'll discuss how to access these in an early lecture.)

  Other recommended books:

  • Bioinformatics and Functional Genomics , by Jonathan Pevsner. A readable introduction to the field. Aimed primarily at biologists, provides somewhat less detail than Mount, but is slightly more approachable.
  • The Cartoon Guide to Genetics by Larry Gonick and Mark Wheelis. A surprisingly good and serious introduction to the biological concepts covered in this course.
  • An Introduction to Bioinformatics Algorithms, by N. Jones and P. Pevzner. A new algorithms text focusing on examples motivated by computational biology. Helpful if you've never taken an algorithms class; provides a more gentle introduction to selected topics than the following book.
  • Introduction to Algorithms, by T. Cormen, C. Leiserson, R. Rivest, and C. Stein. The cannonical algorithms textbook. Has nothing to do with biology, but should be on every computer scientist's bookshelf.
  • Introduction to Computational Molecular Biology, by J. Setubal and J. Meidanis. A detailed text focused on computational biology algorithms, aimed at computer scientists, from 1997.
  • Biological Sequence Analysis, by R. Durbin, S. Eddy, A. Krogh, and G. Mitchison. A good computational biology text focusing on sequence analysis, HMMs, and phylogeny. Includes an excellent whirlwind introduction to statistics.
  • Molecular Biology, by David Freifelder. A general introductory molecular biology text. Easy to read, a gentle introduction to the topic.
  • Molecular Biology of the Gene, by J. Watson, N. Hopkins, J. Roberts, J. Steitz, and Alan Weiner. A more advanced and detailed molecular biology text. A very thorough index makes this a good reference book.

• Computational resources: You will need access to a computer with an internet connection and support for whatever programming language / tools you intend to use. If you need help in obtaining resources for this purpose, please contact the instructor or teaching assistant as soon as possible.

• Term paper or project: Students are expected to do a substantial piece of work studying one area of computational biology in more depth than we can cover in class. There are two options for this requirement: a review paper, or a project. I expect most students new to the field to choose to write a review paper to explore one area or topic in more detail. However, if you have a particular research area of interest or a project in mind of the appropriate scope, you are welcome to pursue that instead.

  The typical review paper will involve selecting a topic and writing a thorough survey and critical analysis that illustrates your depth of understanding of that topic. The paper should present the details of the topic as if you were teaching the material to your classmates, and indeed, you will then present some of this material to the class at the end of the term. The paper should cite appropriate primary literature, and should include clear original figures and/or tables to help explain the material. Typical papers will be 5-10 pages in length, excluding the bibliography. You are not required to do any original research or programming for this option, though you are welcome to include a section on future work that outlines a project you might wish to pursue. However, a critical analysis means that you should include your own thoughts on the material you are reviewing. Of the different methods, which ones are best in what circumstances, and why? What do you think is still lacking? What are the next steps in applying these methods? Do you agree with everything the authors wrote about the relevance of their methods? How might the remaining problems be approached?

  Possible topics include (but are not limited to):

  • whole-genome sequence alignment methods
  • sequencing by hybridization
  • genomic mapping methods: hybridization mapping, restriction mapping
  • HMM methods for classifying proteins
  • bi-clustering algorithms for microarray data
  • statistical tools for detecting differential expression
  • promoter sequence identification methods
  • learning dynamic gene regulatory networks
  • copy number variation and its relation to disease
  • computational methods for linking microRNAs to function
  • computational methods for finding disease genes
  • methods for predicting epidemics
  • graphics or image analysis problems in some aspect of biology
  • database integration issues for biological or clinical data

  Choosing a term project is most appropriate if you already know something about the topic you want to pursue, or if you start to study an area and find that you can think of a better way to solve the problem than anyone has yet designed. Generally, projects will involve either writing a program that would be useful for a biomedical research project, designing and testing a new method of solving an existing problem, or testing a biological hypothesis by a novel data analysis. In the past, successful term projects have included such topics as:

  • using statistics to identify errors in DNA sequence
  • building a database containing public data on structural proteins
  • extending an existing theoretical model of transcriptional regulatory networks
  • implementing and evaluating a new method of analyzing gene expression or proteomic data
  • prioritizing lists of genes for biological screening by mining gene expression databases
  • natural language processing to find protein interactions
  • graph-theoretic evaluation of protein interaction networks

  This project must be accompanied by a written report describing the problem, the solution or the results obtained, and the implications of the work. Because the term project usually spans only a few weeks, a description of future work will be an important component of the written report, whether or not you actually choose to pursue the project any further.

Policies

• Grading: Grades will be based on homework assignments (50%), two in-class quizzes (10% each), a term paper or project (20%), and an in-class presentation of that project (10%).

• Late policy: Anything handed in after the start of the class period in which it is due is marked down 10% for each day it is late. (For simplicity, the "first day" is deemed to run from the beginning of the class in which the assignment is due until midnight the following day; subsequent days are counted from midnight to midnight.)

• Collaboration Policy: All written work and code submitted should be your own unless you obtain prior permission to collaborate. You are free to discuss assignments with others in the class unless specifically asked not to, but you must write up your answers and code yourself.

  All sources used should be cited. In other words, if you discuss a homework problem with a classmate, you should list that classmate as one of your references for that problem. A special note about finding solutions on the web: be warned that not everything you read online is correct. Even data from supposedly reputable sources, such as slides posted by faculty at this or other universities, may not have been reviewed by an editor and might contain crucial typos. For this reason, I'd like to discourage you from using Google to tackle the problem sets, but if you choose to do so, you must cite the URL(s) that you used. Directly copying text from any source without attribution is plagiarism and will be dealt with accordingly.

Course Materials

Internet Resources for Computational Biology and Bioinformatics

• NCBI: An excellent starting point for all sorts of genomic data. Includes the next four listings below.
  
• Entrez Gene: The best place to start for gene-specific information. Links to other gene information: sequences, chromosomal locations, description of function, homology to other species.
  
• Entrez: Links to many databases: genes, protein structures, population studies.
  
• OMIM: Online Mendelian Inheritance in Man. Detailed, curated text reviews about diseases and the genes associated with them. Good bibliographic sources for more references and information.
  
• PubMed: Literature search engine. Gets new references in most biomedical literature within a month of publication. Often includes links to online copies of papers (when available).
  
• EBI: the European Bioinformatics Institute. Includes many sequence analysis tools, access to Array Express (a collection of public gene expression data), plus genome browsers, sequence and structure queries, UniProt, and Integr8.
  
• Integr8: browser for information about completed genomes and proteomes.
• FASTA for sequence similarity and homology search.
  
• BLAST for sequence similarity and homology search.
  

Genome Browsers / Viewers:

• UCSC A good browser from the University of California at Santa Cruz.
  
• NCBI map viewer
  
• Ensembl: From the European Molecular Biology Laboratory (EMBL); the European counterpart to NCBI. Perhaps a slightly more user-friendly interface than NCBI genome browser. Also has gene structure information.
  
• AceView: An alternative map viewer good at showing gene structure.

• Human (Homo sapiens) : OMIM, Genes and Disease
• Mouse (Mus musculus) : The Jax (Jackson Laboratories)
• Rat (Rattus norvegicus) : Rat genome resource page
• Fruit fly (Drosophila melanogaster) : BDGP: (Berkeley Drosophila Genome Project)
• Worm (Caenorhabditis elegans) : Wormbase
• Yeast (Saccharomyces cerevisiae) : SaccDB

• PFAM database of protein families and domains.
  
• Integr8: broswer for information about completed genomes and proteomes.
  
• Protein Data Bank (PDB) of solved protein structures.
  
• UniProt, an integrated resource of protein annotation combining data from SWISSPROT, TrEMBL, and the PIR.
  
• SCOP a hierarchal classifier of solved protein structures
  
• CATH another hierarchical classifier of solved protein structures.
  

Sources of papers:

You can generally get access to these resources (often online) and others through Tisch library or the Health Sciences Library. Bibliographic search tools such as PubMed, CiteSeer, MedMiner, PubGene and Chilibot are also your friends... Consider signing up for some of the table of contents alert services, too.

Note that online access to some journals may be limited to computers within the tufts.edu domain. You can get at these from other domains by going through the Tisch Library page to its electronic journals page, searching for the journal of choice, and then typing in your Tufts ID number.

Conferences

You can find a useful calendar of bioinformatics events at the ISCB website. [By the way, I encourage you to join the ISCB -- it's quite cheap for grad students and cost-effective if you go to one conference or subscribe to one journal.]

• Intelligent Systems for Molecular Biology (ISMB)
• Pacific Symposium on Biocomputing (PSB, "The Hawaii Meeting")
• Research in Computational Biology (RECOMB). Recent Proceedings are in the book stacks in Tisch Library.
• The Genome Informatics Workshop (GIW, "The Japan Meeting")
• European Conference on Computational Biology (ECCB, "The European Meeting")
• The American Medical Informatics Association meeting (AMIA)
• Rocky Mountain Regional Meeting on Bioinformatics (Rocky)

• Bioinformatics
• Journal of Biomedical Informatics
• BioMed Central (BMC) Bioinformatics
• Computational Biology and Chemistry
• Journal of the American Medical Informatics Association
• Journal of Computational Biology
• IEEE/ACM Transactions on Computational Biology and Bioinformatics

An incomplete list of other journals that often contain bioinformatics papers

• Science
• Nature
• Nature Biotechnology
• Nature Genetics
• Nature Medicine,
• Nature Structural and Molecular Biology
• Nature Reviews... (Cancer, Drug Discovery, Genetics, etc.)
• Genome Research
• Genome Biology
• Genomics
• BMC Genomics
• Nucleic Acids Research
• Journal of Biology
• Public Library of Science (PLoS) Biology
• PNAS (Proceedings of the National Academy of Sciences)
• Journal of Machine Learning Research
• Journal of Pharmacogenomics