Protein Prediction with Neural Networks!

1
Protein Prediction with Neural Networks!

• Chris Alvino
• CS152 Fall 06
• Prof. Keller

2
Introduction

• Proteins, made from amino acids
• Polar forces interact for craaazzzy combinatoric
  explosion!
• Just how crazzzzyyy?

3
Real Crazy

• Using crude workload estimates for a
  petaflop/second capacity machine leads to an
  estimate of THREE YEARS to simulate 100
  MICROSECONDS of protein folding.

4
Why Neural Nets?

• Not so crazy
• Relatively accurate results
• 70-80 accurate
• Patterns learned can lead to useful biological
  data
• Used to quickly check existing databases

5
Early Methods Black Box Approach

• Protein Folding Analysis by an Artifical Neural
  Network Approach
• Authors R. Sacile and C. Ruggiero
• Published 1993

6
Early Methods Black Box Approach

• Standard Back Prop Algorithm

7
Early Methods Black Box Approach

• 3 Layers
• Input Window size 13 amino acids
• Hidden Layer 20 neurons
• Output Layer 3 possible (alpha, beta, coil)

8
Early Methods Black Box Approach

• 7 training sets
• Each consists of around 1500 residuals (amino
  acids)
• Training took 3-4 hours

9
Results
10
Artificial Neural Networks and Hidden Markov
Models forPredicting the Protein Structures The
Secondary StructurePrediction in
CaspasesThimmappa S. Anekonda(2002)
11
Current State of the Art

• Neural Networks and Hidden Markov Models

12
Hidden Markov what?

• Hidden Markov models (HMMs), originally developed
  for other applications such as speech
  recognition, are generative, probabilistic models
  of sequential information.
• An observed sequence is modeled as being the
  stochastic result of an underlying unobserved
  random walk through the hidden states of the
  model.
• The parameters of an HMM are the transition
  probabilities between the hidden states and the
  symbol emission probabilities from each hidden
  state.

13

• State transitions in a hidden Markov model
  (example)x hidden statesy observable
  outputsa transition probabilitiesb output
  probabilities

14
Caspases, the friendly Ghost

• Caspases are a family of intracellular cysteine
  endopeptidases.
• They play a key role in inflammation and
  mammalian apoptosis or programmed cell death.

15
Clash of the Titans

• PHDSec
• Utilizes evolutionary information
• PSIPRED
• Uses iterated PSI-BLAST profiles as input instead
  of multiple sequeence alignments like PHDSec
• SAM-T02
• Uses ANN and HMM
• PROF King
• Uses seven GOR-based predictions and ANN

16
(No Transcript)