Data and Text Mining for Computational Biology

1
Data and Text Mining for Computational Biology

• Introduction

2
Course information

• CS 6365
• Data and Text Mining for Computational Biology
• Meets Tuesday and Thursday 700-815
  pm at ECSS 2.412

3
Instructor

• Vasileios Hatzivassiloglou
• Associate Professor, Computer Science
• Founding Professor, Bioengineering
• Research focus Discover knowledge from massive
  amounts of raw data
• data not the same as information
• information overload

4
Research Interests

• Text analysis, machine learning, intelligent
  information retrieval, summarization, question
  answering, bioinformatics, medical informatics

5
Contact information

• Office hours Tuesday and Thursday 600-700pm
  and by appointment
• Office location ECSS 3.406
• vh_at_hlt.utdallas.edu
• [phone]
• Teaching Assistant TBA

6
Course goals

• Introduce the field of bioinformatics
• Discuss primary techniques used for data mining
• Introduce text mining and additional issues it
  brings to data mining methods
• Use examples from computational biology

7
Intended audience

• For both computer scientists and biologists
• Not an easy task to balance the two
• Focus on data and text mining algorithms and
  applications
• Coverage of machine learning background
• No extensive algorithmic analysis / computational
  complexity
• Medium level of programming

8
Prerequisites

• Officially CS 6325 Introduction to
  Bioinformatics
• Waived for this offering of the course
• You should know
• Basic data structures (multidimensional arrays,
  hash tables, binary trees)
• One high-level programming language and be able
  to adapt to a new one as needed
• Be able to install and use external software
  packages

9
You need not know

• Molecular biology
• Machine learning
• Data mining (in general)
• Text analysis / natural language processing
• Information retrieval
• Artificial intelligence

10
Course level

• Introductory graduate course (MS or first-year
  PhD)
• Maturity in programming and data structures as of
  a Computer Science senior
• Ability (and interest in) accessing the primary
  literature in a guided fashion

11
Course structure

• 6 lectures on biological background and
  bioinformatics in general
• 6 lectures on data similarity
• 8 lectures on data mining methods
• 3 lectures on text mining and knowledge mining
  methods
• student presentations of research papers (3-4
  sessions)

12
Expected work load

• Two homework sets given in mid-to-late September
  and mid-to-late October
• Two weeks to turn in each homework set
• Mid-term exam in early October
• Each student selects two or three research papers
  to review in late October
• Student presentations of research papers in the
  last week of November / first week of December
• Final exam

13
Course project

• In lieu of the research papers and presentation,
  students may elect to work on a project in teams
  of two or three
• Project is chosen by the students with the advice
  and consent of the instructor
• Project investigation/implementation should be
  approximately 1.5-2 times the work required for a
  regular homework

14
Programming

• Each student selects their own programming
  language (must be available at UTD and accessible
  to TA)
• Examples C, C, Java, Perl, Python
• Can also use a package/programming environment
  specifically tailored to bioinformatics

15
One likely package

• R (http//www.r-project.org/)
• R is the free alternative to S-Plus developed at
  ATT research
• S-Plus is the extensible, programmable
  alternative to statistical packages like SAS and
  SPSS
• If you know C, you will be right at home with R

16
Another likely package

• BioPerl (http//bio.perl.org/)
• A collection of library modules in Perl written
  by and for bioinformaticians
• Perl supports high-level operations such as
  hashes as a basic data structure, string
  matching, and regular expressions
• Perl is really bad at OOP and efficiency
• Easy to learn

17
Grading

• Class participation 20
• Homework assignments 30 (total)
• Midterm 10
• Research paper presentation or project 20
• Final exam 20

18
Textbooks

• No good integrated textbook on data mining from a
  computational biology perspective
• We will use a text book covering bioinformatics
  algorithms and another text book on data mining
  in general, and additional chapters from other
  books and research articles
• Copies of chapters / research articles will be
  provided

19
Recommended textbook 1

• An Introduction to Bioinformatics Algorithms
  (Computational Molecular Biology), by Neil C.
  Jones and Pavel A. Pevzner, MIT Press, 2004.
• ISBN 0262101068
• 448 pages
• Available on Amazon.com for 41, Barnes and Noble
  for 60

20
Recommended textbook 2

• Data Mining Concepts and Techniques by Jiawei
  Han and Micheline Kamber, Elsevier, second
  edition, 2006.
• ISBN 1558609016
• 800 pages
• Available on Amazon.com for 52, Barnes and Noble
  for 65

21
Supplementary textbooks

• Bioinformatics The Machine Learning Approach
  by Pierre Baldi and Soren Brunak, 2nd edition,
  2001.
• Data mining multimedia, soft computing, and
  bioinformatics by Sushmita Mitra and Tinku
  Acharya, 2003.
• Both of the above are available as full-text
  eBooks via http//library.utdallas.edu.

22
Background reading

• Biology Molecular Biology of the Cell by Bruce
  Alberts et al., 4th edition, 2002.
• Machine learning Machine Learning by Tom
  Mitchell, 1997.

23
Background reading (II)

• Statistics The elements of statistical
  learning data mining, inference, and prediction
  by Trevor Hastie, Robert Tibshirani and Jerome
  Friedman, 2001.
• Data structures and algorithms Introduction to
  Algorithms, by Thomas H. Cormen, Charles E.
  Leiserson, Ronald L. Rivest, and Clifford Stein,
  2nd edition, 2001.

24
So what is it all about?

• Three parts
• Bioinformatics / computational biology
• Data mining
• Text mining

25
Bioinformatics

• A fast developing discipline
• We will discuss
• basic concepts of molecular biology
• databases of biological data
• structure and function of DNA, RNA, proteins
• sequence searching (BLAST)
• sequence similarity and comparison
• protein structure (2D and 3D)
• protein motifs and patterns
• microarrays
• phylogenetics

26
Data mining

• Given a large amount of data of known types,
  extract useful information
• We will discuss
• data cleanup and outliers
• model construction
• data and dimensionality reduction
• classification
• prediction / probability estimation
• clustering
• measuring performance

27
Text mining

• Not only we have a large amount of raw data, but
  we dont know what each item means
• We will discuss
• tokenization and basics of text processing
• recognition of terms and entities
• classification
• dictionary creation
• relationship learning and extraction
• document level clustering and information
  retrieval