Algorithms in Bioinformatics: A Practical Introduction

1
Algorithms in Bioinformatics A Practical
Introduction

• Project
• Motif finding using ChIP-seq peak data

2
Transcriptional Control (I)
3
Transcriptional Control (II)
TATAAT is the motif!
4
Motif model
TTGACA TCGACA TTGACA TTGAAA ATGACA TTGACA GTGACA T
TGACT TTGACC TTGACA
Consensus Pattern
TTGACA
Positional Weight Matrix (PWM)

• Motif can be described in two ways based on the
  binding sites discovered

5
ChIP experiment

• Chromatin immunoprecipitation experiment
• Detect the interaction between protein
  (transcription factor) and DNA.

6
Peak data

• Peak data represents the locations where a
  particular TF binding.
• The data tells us the locations and intensities.
• (Note that due to experimental error, peaks of
  low intensity may be noise.)

ChIP-seq data for Human (MCF7) E2 treatment at
45min
chr1883,686-958,485
7
Our aim

• Given the DNA sequences of those peaks, find
  motifs which occur in those peak regions.
• For the example below, we have two motifs TTGACA
  and GCATC.
• Note that each instance has at most 1 mutation.

GCACGCGGTATCGTTAGCTTGACAATGAAGAATCCCCCCGCTCGACAGT
GCATACTTTGACACTGACTTCGCTTCTTTAATGTTTAATGAAACATGCG
CCCTCTGGAAATTAGTGCGGCATCTCACAACCCGAGGAATGACCAAATG
GTATTGAAAGTAAGGCAACGGTGATCCCCATGACACCAAAGATGCTAAG
CAACGCTCAGGCAACGTTGACAGGTGACACGTTGACTGCGGCCTCCTGC
GTCTCTTGACCGCTTAATCCTAAAGGCCTCCTATTAGTATCCGCAATGT
GAACAGGAGCGCGAGCCATCAATTGAAGCGAAGTTGACACCTAATAACT
8
Input (I)

• From every peak, we get approximately /-200 DNA
  sequence
• gtcmyc_1_chr1_4842133_4842148_range_chr1_4841934_48
  42348_intensity_20
• CCTCCATACCAGCCCCAATGTTCTGCGTTCCCGAATGAAAGACACACAAC
  ACAGCCTTTATATTTTGATATGCCTAAAACTGCTCAATGGCTGGGCCACT
  TCCTAGCTAGTATCCACGTGGCTATCCCACCTCTCTCTGATATTCCCAAG
  TCATTACTTACTAAAATCTGTAATTACATCTTTGCTGCCCTAGGCCCAAT
  CTGGCAGCCCTCCTGTGGCCCCTCAGGCTACTACATGGCAGCTAAGCTCT
  CTGACCCACATCTTCTCAGGCACCGTGCCTCCTCTTCTCCACCTTATTCA
  AACATGGTGGCTCTCCTTCCTCCTTCTTCCTGTCTGTCCCCAGCCTGGGA
  ATTCTAAAAGTCCCACCTCTGTCTGCCCTGTTCAGCCATTGGCTGTCGGC
  ATCTTTATTTACGAG
• gtcmyc_2_chr1_5073201_5073215_range_chr1_5073002_50
  73415_intensity_15
• GGTCATAAACCAAGCTTCTTCAAAGATTTTTGGCTTTTTGGCACCAGTGG
  CCTGCAGGGTGGCGAGCTCTGCCAGTTTGAAGTGACCAAGTTAAGTGGCC
  TGGGAAAGGCCATTTGGTGCGCGGTCCAGCAGTTTTGGGCGCTCTCGGCT
  TCCGCCCTCAGCTGCGGTCACGTGCGGCTGCTCACGTGCCAGACGCTGCT
  GTCACTTCGTAGCTGTTCCGGCTTCCTCTGAGTGAGGCTCGCAACGTCTC
  CCACGGAGTCGCCTTCGTTCTGCTCTGGGTCTCCCGTGGCCACTGAGACC
  TCGGAGCTCGACCGGCGCCTGCCCGCCCGTGCGGCCCTCACTCCCCGAGG
  CTATCCAGGTGAGGCCGCCTGGGGTCCCTCCCCGGCTCCGGAGAGCCGAC
  TGGTTTCCCTGCCG
• gtcmyc_3_chr1_9530642_9530652_range_chr1_9530443_95
  30852_intensity_36
• GTAGTCCCAACCAGGTCCTGAGCTGGTTAGCCAACCCTCAGCGCCAGTCG
  GGCCAACATCCGGTGACGAATCCAAGTCCCGCCTCTAAGCCCATCTGCTG
  TCCAATGCCGCCCTCTGCCGGTCTTTACCTCCCCGCCTAGCTGTGAGCCG
  CTTCCAGACAACCCGGAAGTGATCTTTCCTCTTCCGGATTACGGGTCCGG
  ACGTCCGCACGTGGTTGCCGGTTTAGGGTGCTGCTGTAGTGGCGATACGT
  CCCGCCGCTGTCCCGAAGTGAGGGATCCGAGCCGCAGCGAGAGCCATGGA
  GGGCCAGCGCGTGGAGGAGCTGCTGGCCAAGGCAGAGCAGGAGGAGGCGG
  AGAAGCTGCAGCGCATCACGGTGCACAAGGAGCTGGAGCTGGAGTTCGAC
  CTGGGCAACC

9
Input (II)

• A set of sequences which are likely containing no
  motif.
• gtSEQ_1
• AACAAGGGAAAGAGTAGTGAGTGCTTCTTTCTATTCAGAGGGAGGGGAAG
  TTGCTGTTAGCTAAGACAGTCAGGACTGAGAAGGGGGGGGGGGGTTTAAC
  TCTCCTGGAGGGAGCTGAGAGGTAAAGGGAGGGGCGTGAGGTAGAACAAG
  CCGAGAACACAGGGCAGGTTGGTCTGACTCCAGAGCACAGTGCAGGAGCC
  CGGAAGTTGACTCAGTTCAGTTAGCAAGTATTTTCACACAAGGCGTGAAC
  ACTGAAGACAAAAGCAAGAGACACAGCTCTATCTCTAAGAAGATTTTCAG
  AGCCAAGATCGATGGGGCACACCTGTTAATCCCAGCACTTAGGAGGCTGA
  GGCAGGAGGATCCCAAGTTCAAAACCAGCCTGGACTTGTTTTAAGGAAAA
  
• gtSEQ_2
• AAAAAAAAAAAAAAGACTTCCAGTTTAATAAATGACCAATTCAGGAATGG
  AGATTAGGGCTGGATGACAAGTTTTTAATTGTCAAGGACTCAATTCTGTT
  TATCAGTTGGTATGGAATTATGTAAGCTTTTAGCGATATGACCGCACGGA
  GCAGTGTAGAGAGTGATCTGAGAGACGCTTGGGGGTCAGGATGGAGATAG
  AACTCCCTCTCTATTAGAAGGTGTTTGGTGGTAGGTAACCCTGGGCTAGC
  ATGGTGGGTCTCTTCTTACTTAGGCTTCCATCTTTGTGGTTCAAATCCAA
  GAAGGACCTGCGTTCCCTCCCTCCTTGTGATCAGCTGATTGCTAGAGCAT
  AACTCATCTTAACTTCTCATGTACTCTCCGGGTACAGGAAGGGAGGGGGC
  
• gtSEQ_3
• CCACTGCTGACAGTGGAGCATGAAACGACCGGCTTCCTGACTATGTTGGT
  ACCCTTTCAGGAGCCTAAAACAGTGCTTTCAATACTTGTGTCTATGTCTG
  TTAGCCACAACTTTCTAGTTTCCCAGAGAGATTTTGAAGTGTAGTTTTGT
  ATTTGCTCAAATATATATTCATATGGTGAGGTGCACATTTTTTATATTAT
  ATTTTTATTCATTTATTTTTGGTGCTTGGGAATTATACTCTAGGAATAAA
  GCGCCTGGTAGAAAGTGGCACACATCTTTAATCCCAGCACTCAGGAAGCA
  GAGGCAGACAAATCTCTGCGTTCCAGGACAGCCTGGTCTATAGAGCAAGG
  TCCAAGCCAGCCAGGTTTACACAAAGAAACCTAGTGTGGAAAAGACAAAA
  

10
Output

• You need to output a list of candidate (ranked)
  motifs.
• You can model the motif as PWM or consensus
  sequence.
• If you model the motif as a PWM, one of the
  answer for the previous dataset is
• You may also return other significant motifs.

11
Aim of the project

• Given a sample file and a background file,
• you need to implement a method which output a
  list of motifs.
• You need to take advantage of the fact that this
  is a ChIP-seq dataset
• Hint Read papers on ChIP-seq and understand its
  properties.