Transcript StuartBrown-Teaching

Teaching Bioinformatics to
Undergraduates
http://www.med.nyu.edu/rcr/ASM
Stuart M. Brown
Research Computing, NYU School of Medicine
I. What is Bioinformatics?
II. Challenges of teaching
bioinformatics to undergraduates
III. Common bioinformatics tools
that you can use for teaching
IV. The limits of knowledge
V. Resources for the teacher
I. What is Bioinformatics?




The use of information technology to collect, analyze,
and interpret biological data.
The use of software tools that deal with biological
sequences, genome analysis, molecular structures,
gene expression, regulatory and metabolic modeling
Computational biology - the design of new algorithms
and software to support biology research
The routine use of computers in all phases of biology
and medicine
The Human Genome Project
A Genome Revolution in Biology
and Medicine



We are in the midst of a "Golden Era" of
biology
The Human Genome Project has produced a
huge storehouse of data that will be used to
change every aspect of biological research
and medicine
The revolution is about treating biology as an
information science, not about specific
biochemical technologies.
The job of the biologist is changing

As more biological information becomes
available …and laboratory equipment becomes
more automated ...
– The biologist will spend more time using computers
– The biologist will spend more time on experimental
design and data analysis (and less time doing
tedious lab biochemistry)
– Biology will become a more quantitative science
(think how the periodic table affected chemistry)
II. Why teach bioinformatics in
undergraduate education?

Demand for trained graduates from the biomedical
industry

Bioinformatics is essential to understand current
developments in all fields of biology

We need to educate an entire new generation of
scientists, health care workers, etc.
Use bioinformatics to enhance the teaching of other
subjects: genetics, evolution, biochemistry

Biochemistry & Protein Structures

"Hands-on graphics is a powerful enhancement to
learning, particularly individualized learning. There
is powerful synergy in learning about proteins and
learning simultaneously about how to represent and
manipulate them with computer graphics. When
students learn to use graphics they see proteins and
other complex biomolecules in a new and vivid way,
and discover personal solutions to the problem of
"seeing" new structural concepts."
Molecular Graphics Manifesto
Gale Rhodes
Chemistry Department, Univ. of Southern Maine
Challenges of presenting
bioinformatics to undergraduates




Requires a deep understanding of molecular
biology - lots of prerequisites
Training users or makers of these tools?
A good bioinformatics program will require
substantially more math and statistics than
most existing molecular biology and
computer science curricula.
Who will teach?
Different Programs,
Different Goals




Integrate into existing biology courses:
– genetics, molecular biology, microbiology
Make one or a few cross-disciplinary courses
– jointly taught by biology and computing faculty
– open to both biology and computing students
Create a curriculum for a true bioinformatics major
(is this a double major?)
Are you training for employment or providing the
fundamentals for advanced training?
Shallow End


This workshop will focus on faculty skills
needed at the shallow end of the continuum
(a few lectures or a short course).
Use bioinformatics to teach biological
concepts
» Evolution
» Genetics
» Protein structure and function
How much Computing skills?


Bioinformatics can be seen as a tool that the
biologist needs to use - like PCR
Or should biologists be able to write their own
programs and build databases?
– it is a big advantage to be able to design exactly the
tool that you want
– this may be the wave of the future

Is your school going to train "bioinformatics
professionals" or biologists with informatics
skills?"
Designing a Curriculum



To really master bioinformatics, students need
to learn a lot of molecular biology and genetics
as well as become competent programmers.
Then they need to learn specific bioinformatics
skills - dealing with sequence databases,
similarity algorithms, etc.
How can students learn this much material and
still manage a well rounded education?
– Graduates of these programs will become scientists
and managers. Writing and presentation skills are
essential components of their education.
Different Schools have
Different Biases

There are still only a handful of
bioinformatics undergraduate programs
– [Many more schools offer a single course or a "specialized
track" similar to a biotechnology major]

You can generally predict the bias according
to what school/department hosts the program
– Computer Science vs. biology
– Biomedical engineering
– Medical informatics (library science)
Teaching the Teachers




There are more graduate level bioinformatics
programs, but they are all very new.
Graduates of these programs will have many
opportunities as more schools gear up to offer
bioinformatics training
The reality is that most schools will draft
existing faculty - often jointly from Bio and
CompSci departments
We need to train an entire generation of
existing faculty in a new discipline
Teaching Tips

Strike a balance between theory and practical
experience
– early bioinformatics training should be about what
you can do with the tools
– deeper training can focus on how they work

Balance the "click here" tutorials against letting
them figure it out for themselves
– it will be different when they look at it next time
– real bioinformatics work involves finding ways to
overcome frustrations with balky computer systems
Training "computer savvy"
scientists

Know the right tool for the job

Get the job done with tools available

Network connection is the lifeline of
the scientist

Jobs change, computers change,
projects change, scientists need to be
adaptable
III. Bioinformatics Tools
You Can Use









GenBank - genes, proteins, genomes
Similarity Search tools: BLAST
Alignment: CLUSTAL
Protein families: Pfam, ProDom
Protein Structures: PDB, RasMol
Whole Genomes: UCSC, Entrez Genomes
Human Mutations: OMIM
Biochemical Pathways: KEGG
Integrated tools: Biology Workbench,
BCM SearchLauncher
Large Databases

Once upon a time, GenBank sent out
sequence updates on CD-ROM disks a few
times per year.

Now GenBank is over 40 Gigabytes
(11 billion bases)

Most biocomputing sites update their copy of
GenBank every day over the internet.

Scientists access GenBank directly over the
Web
Finding Genes in GenBank
•These billions of G, A, T, and C letters
would be almost useless without
descriptions of what genes they contain,
the organisms they come from, etc.
•All of this information is contained in the
"annotation" part of each sequence record.
Entrez is a Tool for Finding Sequences

GenBank is managed by the NCBI (National Center
for Biotechnology Information) which is a part of
the US National Library of Medicine.

NCBI has created a Web-based tool called Entrez
for finding sequences in GenBank.
http://www.ncbi.nlm.nih.gov

Each sequence in GenBank has a unique “accession
number”.

Entrez can also search for keywords such as gene
names, protein names, and the names of orgainisms
or biological functions
Entrez Databases contain more than
just DNA & protein sequences
Type in a Query term

Enter your search words in the
query box and hit the “Go” button
Refine the Query




Often a search finds too many (or too few) sequences, so
you can go back and try again with more (or fewer)
keywords in your query
The “History” feature allows you to combine any of your
past queries.
The “Limits” feature allows you to limit a query to specific
organisms, sequences submitted during a specific period of
time, etc.
[Many other features are designed to search for literature in
MEDLINE]
Related Items
You can search for a text term in sequence annotations or in
MEDLINE abstracts, and find all articles, DNA, and protein
sequences that mention that term.
Then from any article or sequence, you can move to "related
articles" or "related sequences".
•Relationships between sequences are computed with BLAST
•Relationships between articles are computed with "MESH" terms
(shared keywords
•Relationships between DNA and protein sequences rely on accession
numbers
•Relationships between sequences and MEDLINE articles rely on both
shared keywords and the mention of accession numbers in the articles.
Database Search Strategies


General search principles - not limited to
sequence (or to biology)
Use accession numbers whenever possible
Start with broad keywords and narrow the
search using more specific terms
Try variants of spelling, numbers, etc.
Search all relevant databases

Be persistent!!



>gb|BE588357.1|BE588357 194087 BARC 5BOV Bos taurus cDNA 5'.
Length = 369
Score = 272 bits (137), Expect = 4e-71
Identities = 258/297 (86%), Gaps = 1/297 (0%)
Strand = Plus / Plus
Query: 17
Sbjct: 1
Query: 77
Sbjct: 60
aggatccaacgtcgctccagctgctcttgacgactccacagataccccgaagccatggca 76
|||||||||||||||| | ||| | ||| || ||| | |||| ||||| |||||||||
aggatccaacgtcgctgcggctacccttaaccact-cgcagaccccccgcagccatggcc 59
agcaagggcttgcaggacctgaagcaacaggtggaggggaccgcccaggaagccgtgtca 136
|||||||||||||||||||||||| | || ||||||||| | ||||||||||| ||| ||
agcaagggcttgcaggacctgaagaagcaagtggagggggcggcccaggaagcggtgaca 119
Query: 137 gcggccggagcggcagctcagcaagtggtggaccaggccacagaggcggggcagaaagcc 196
|||||||| | || | ||||||||||||||| ||||||||||| || ||||||||||||
Sbjct: 120 tcggccggaacagcggttcagcaagtggtggatcaggccacagaagcagggcagaaagcc 179
Query: 197 atggaccagctggccaagaccacccaggaaaccatcgacaagactgctaaccaggcctct 256
||||||||| | |||||||| |||||||||||||||||| ||||||||||||||||||||
Sbjct: 180 atggaccaggttgccaagactacccaggaaaccatcgaccagactgctaaccaggcctct 239
Query: 257 gacaccttctctgggattgggaaaaaattcggcctcctgaaatgacagcagggagac 313
|| || ||||| || ||||||||||| | |||||||||||||||||| ||||||||
Sbjct: 240 gagactttctcgggttttgggaaaaaacttggcctcctgaaatgacagaagggagac 296
Sample Multiple Alignment
Protein domains
(from ProDom database)
Limits on best Matched Annotation Inheritance
result from many things including multi domain proteins transitivity.
New sequence
Closest database annotated entry
Original studied protein from which
annotation was inherited.
Protein Structure



It is not really possible to predict protein
structure from just amino acid sequence
PDB is a database of know protein structures
(determined by X-ray crystallography and
NMR)
There are also very handy structure viewers
such as RasMol that are free for any
computer
Genome Browsers

Scientists need to work with a lot of
layers of information about the genome
– coding sequence of known genes and
cDNAs
– computer-predicted genes
– genetic maps (known mutations and
markers)
– gene expression
– cross species homology
UCSC
Ensembl at EBI/EMBL
Human Alleles

The OMIM (Online Mendelian Inheritance
in Man) database at the NCBI tracks all human
mutations with known phenotypes.

It contains a total of about 2,000 genetic
diseases [and another ~11,000 genetic loci with
known phenotypes - but not necessarily known gene
sequences]

It is designed for use by physicians:
– can search by disease name
– contains summaries from clinical studies
KEGG: Kyoto Encylopedia of
Genes and Genomes



Enzymatic and regulatory pathways
Mapped out by EC number and crossreferenced to genes in all known organisms
(wherever sequence information exits)
Parallel maps of regulatory pathways
Integrated Online Tools

National Database/Sequence Analsysis Servers:
– NCBI, EMBL/EBI, DDBJ

Tools for specific types of data or problems
– Expasy (Protein, Mass Spec, 2-D PAGE)
– 3-D Protein Structures:
PDB, Predict Protein Server

Education oriented tools
– Biology Workbench

Collections of links to other servers
– BCM SearchLauncher
The Limits of our Knowledge

Bioinformatics is a very dynamic discipline
– Teachers can't know everything in the field


The databases are clumsily built
Biology is vastly more complex than our
software
– Lots of our current bioinformatics programs don't
work well
– We don't have even theoretical solutions for Gene
prediction, alternative splicing, protein structure &
function prediction, regulatory networks
What is a Gene?

For every 2 biologists, you get 3 definitions

“A DNA sequence that encodes a
heritable trait.”
The unit of heredity

Is it an abstract concept, or something you
can isolate in a tube or print on your screen?

“Classic” vs.. “modern” understanding of
molecular biology
Genome Confusion

The sequence of a gene in the genome includes:
–
–
–
–
–

protein coding sequence
introns and exons
5' and 3' untranslated regions on the mRNA
promoter and 5' transcription factor binding sites
enhancers??
What about alternative splicing?
– Multiple cDNAs with different sequences (that
produce different proteins) can be transcribed from
the same genomic locus
V. Teaching Resources

The Biology Student WorkBench
http://peptide.ncsa.uiuc.edu/bioswb.html

RasMol/Chime/Protein Explorer
http://www.umass.edu/microbio/rasmol/
http://www.umass.edu/microbio/chime/
http://www.umass.edu/microbio/chime/explorer/

Bioinformatics.org

Other courses - It ain't cheating to learn from
your peers

Terri Attwood's Web Biocomputing tutorials
http://www.biochem.ucl.ac.uk/bsm/dbbrowser/c32/index.html
http://www.biochem.ucl.ac.uk/bsm/dbbrowser/jj/prefacefrm.html

Sequence Analysis on the Web
– Christian Büschking and Chris Schleiermacher
http://bibiserv.techfak.uni-bielefeld.de/sadr/

Online Lectures on Bioinformatics
–
Hannes Luz, Max Planck Institute for Molecular Genetics
http://lectures.molgen.mpg.de/

Using Computers in Molecular Biology
–
Stuart Brown, NYU School of Medicine
http://www.med.nyu.edu/rcr/rcr/course/index.html
Teach Yourself Bioinformatics on the Web
http://www.med.nyu.edu/rcr/rcr/btr/index.html
Long Term Implications
 A "periodic
table for biology" will lead to
an explosion of research and discoveries we will finally have the tools to start
making systematic analyses of biological
processes (quantitative biology).
Understanding the genome will lead to
the ability to change it - to modify the
characteristics of organisms and people in
a wide variety of ways

Genomics in Medical Education
“The explosion of information about the new
genetics will create a huge problem in
health education. Most physicians in
practice have had not a single hour of
education in genetics and are going to be
severely challenged to pick up this new
technology and run with it."
Francis Collins
Bioinformatics:
A Biologist's
Guide to
Biocomputing
and the
Internet
Stuart M. Brown, Ph.D.
[email protected]
www.med.nyu/rcr