Highresolution computational models of genome binding events

1
High-resolution computational models of genome
binding events

• Yuan (Alan) Qi
• Joint work with Gifford and Young labs

Dana-Farber Cancer Institute Jan 2007
2
ChIP-chip Experiments

• ChIP-chip data
• Encode valuable information about protein-DNA
  binding events.
• Goal
• Decode accurate binding information from the
  noisy data.
• Challenges
• Noise
• Joint influence of multiple binding events

3
Joint Binding Deconvolution
Data Likelihood
Prior Distributions
Hyper Prior Distributions
JBD generative probabilistic graphical model.
4
Shear Distribution
(b) An influence function is derived from the
measured fragment size distribution.
(a) The distribution of DNA fragment sizes
produced in the ChIP protocol were experimentally
measured and statistically modeled.
5
Approximate Bayesian Inference
Exact Bayesian posterior of binding events
Where and
Non-conjugate models, thousands of variables -gt
Intractable calculations of the exact posterior
distribution!
Message passing algorithm (Expectation
propagation)
EP iteratively refines the factor approximations
(i.e., messages) to improve the posterior
approximation.
6
EP in a Nutshell

• Approximate a probability distribution by
  simpler parametric terms
• Each approximation term lives in an
  exponential family (e.g., Gaussian or Gamma
  distributions).

7
EP in a Nutshell

• Three key steps
• Deletion Approximate the leave-one-out
  posterior distribution for the ith factor.
• Minimization Minimize the following KL
  divergence by moment matching.
• Inclusion

8
Results
9
(No Transcript)
10
(No Transcript)
11
Spatial resolution comparison between JBD and
other methods

• The average distance of JBDs Gcn4 binding
  predictions to motif sites is smaller than for
  other methods, and JDB identifies more known Gcn4
  targets.

12
JBD better resolves proximal binding events than
do other methods. Shown here is performance of
the JBD, MPeak and Ratio methods on 200 simulated
DNA regions each containing two binding events.
13
Using binding posterior to guide motif discovery

• Approach
• Using binding posterior probabilities derived
  from the ChIP-chip data to weight sequence
  regions differently for motif discovery.
• Results
• Finding Mig2 motif while a standard motif
  discovery algorithm (e.g., MEME) failed.
• Note that the correct motif for Mig2 was not
  recovered when using the Ratio method to analyze
  the ChIP-chip data.

14
Positional priors for motif discovery improve
robustness to false input DNA sequence regions.
15
Questions?