Transcript Topics in Computational Biology

Topics in Computational Biology
(COSI 230a)
Pengyu Hong
09/02/2005
Background

As high-throughput methods for biological
data generation become more prominent
and the amount and complexity of the
data increase, computational methods
have become essential to biological
research in this post-genome age.
Background
High-throughput methods …
Transcriptional profiling
cDNA arrays
Simultaneously monitor the
transcriptional activities of
tens of thousands of genes.
Oligonucleotide arrays
• Functions of gene
• Relationships between
gene-products
• ……
• New drugs
• Personalized
medicine
• ……
Background
High-throughput methods …
Transcriptional
profiling
High-Content Screening
104 images in one experiment
Background
High-throughput methods …
Transcriptional
profiling
High-Content
Screening
Statistical
Machine
Learning
Score histogram of
phenotype images
Score histogram of
wildtype images
Background
High-throughput methods …
Transcriptional
profiling
High-Content
Screening
……
Publications
PubMed: 15+ million
bibliographic citations
and abstracts
Background

In turn, biological problems are motivating
innovations in computational sciences,
such as computer science, information
science, mathematics, and statistics.
Background
Complex biological systems need novel
computational methods …
Stimuli
S1
K
S2
S3
P1
K2
1
Signal transduction
networks
K3
P3
Transcriptional
regulatory networks
Cellular phenotypes
P2
K4
K5
Gene
group 3
Gene
group 1
Gene
group 2
Gene
group 4
Background
Complex biological systems need novel
computational methods …
Stimuli
S1
K
S2
S3
P1
K2
Spatial
1
Signal transduction
networks
K3
P3
Transcriptional
regulatory networks
Cellular phenotypes
P2
K4
Temporal
K5
Gene
group 3
Gene
group 1
Gene
group 2
Gene
group 4
Background
Large scale data needs novel information systems
Local Data
Local Data
SOAP APIs
Functions
Functions
UBIC2 Unit A
Remote biological
databases
LocusLink
RGD
HGNC
MGI
UCSC
……
UBIC2 Unit B
Ubiquitous bio-information computing (UBIC2)
• Integrate heterogeneous data
Background
Novel Human-computer interfaces (e.g., visualization, multimodal
interaction techniques, and context-aware learning functions.) are
needed to help biologists efficiently navigate through the
complicated landscape of biomedical information and
effectively manipulate various computational tools.
• Collect information while surfing
the Internet.
• Manage multimedia biological
information (text, PDF, images,
sequences, etc.)
GeneNotes
• Functional based literature search
(about to release this year).
Background

There is high demand for scientists who
are capable of bridging these disciplines.
Trend
Shallow biology + Shallow computing
Shallow biology
+
Deep computing
or
Deep biology
+
shallow computing
Deep biology + Deep computing
Background
High demand for interdisciplinary scientists who
are capable of speaking multiple “languages”.
Design
experiment
s
Analyze
data
Generate biologically
meaningful
computational results.
Carry out
experiment
s
Generate informative
experimental data.
Background
High demand for interdisciplinary scientists who
are capable of speaking multiple “languages”.
Design
experiment
s
Analyze
data
Generate biologically
meaningful
computational results.
Carry out
experiment
s
Generate informative
experimental data.
Background
High demand for interdisciplinary scientists who
are capable of speaking multiple “languages”.
Goal: Customize cDNA arrays to
measure the temporal transcriptional
profiles of a set of genes
Design
experiment
s
Analyze
data
Carry out
experiments
Genes besides those of interest?
Computational tools?
How to choose time point for sampling?
Background
High demand for interdisciplinary scientists who
are capable of speaking multiple “languages”.
Design
experiment
s
Analyze
data
Goal: Use a 384 well plate to test the
effects of various treatments on cells.
Carry out
experiments
Duplicates?
Treatment arrangement?
Base line?
Goal

Create an environment


Transcends traditional departmental
boundaries
Facilitates communications between
researchers from life sciences and
computational sciences.
Goal

Learn knowledge (bio + comp) specific to
a set of problems.
• Regulatory motif finding
• Microarray data analysis
• Biomedical literature mining
• Signal transduction network modeling
• Cis-regulatory network discovery
• ……
Goal

Acquire skills


Initiate interdisciplinary collaborations (choose
research partners)
Establish long-term win-win collaborations.
Key: Seek first to understand, then to be
understood. (Stephen R. Covey)
Main Themes

Presentation

Term Project
Main Themes

Presentation

Materials: Your own work or other
people’s published results
 Your
own work: This is a good opportunity
for you to attract collaborators.
 Published papers: Suggest to choose one
and search for related ones.


60 Minutes followed by questions and discussions
Written report after presentation
Main Themes

Presentation



Materials: Your own work or other people’s published
results
60 minutes presentation followed by
questions and discussions
Written report after presentation
Main Themes

Presentation



Materials: Your own work or other people’s published
results
60 minutes presentations followed by questions and
discussions
Written report after presentation
Background of the research
 Motivation for the research
 Approach
 Results
 Criticisms and/or suggestions for improvement.

Main Themes

Term project


Decide by mid-term
Due on 12/22 mid-night.
Evaluation

Grading will be based on class
participation and on the project.
Evaluation


Grading will be based on class
participation and on the project.
Teamwork is strongly encouraged !!!

Indicate the contribution of each individual.
Questions?




Prepare your presentation.
Choose a right project.
……
Me at:



Office hour Tue & Fri 4:30-5:30pm.
Office Volen 135
Email: [email protected].
Please fill the form and return it to me now.
Thanks