Finding sequence motifs in PBM data Workshop Project

1
Finding sequence motifs in PBM data Workshop
Project
Yaron Orenstein October 2010
2
Outline

• 1. Some background again
• 2. The project

3
1. Background

• Slides with Ron Shamir and Chaim Linhart

4
Gene from DNA to protein
Pre-mRNA
Mature mRNA
DNA
protein
transcription
translation
splicing
5
DNA

• DNA a string over the alphabet of 4 bases
  (nucleotides) A, C, G, T
• Resides in chromosomes
• Complementary strands A-T C-G
• Forward/sense strand AACTTGCG
• Reverse-complement/anti-sense strand
  TTGAACGC
• Directional from 5 to 3
• (upstream) AACTTGCGATACTCCTA
  (downstream)

5 end
3 end
6
Gene structure (eukaryotes)
Promoter
DNA
Coding strand
Transcription start site (TSS)
Transcription (RNA polymerase)
Pre-mRNA
Exon
Intron
Exon
Splicing (spliceosome)
5 UTR
3 UTR
Mature mRNA
Stop codon
Start codon
Coding region
Translation (ribosome)
Protein
7
Translation

• Codon - a triplet of bases, codes a specific
  amino acid (except the stop codons) many-to-1
  relation
• Stop codons - signal termination of the protein
  synthesis process

http//ntri.tamuk.edu/cell/ribosomes.html
8
Genome sequences

• Many genomes have been sequences, including those
  of viruses, microbes, plants and animals.
• Human
• 23 pairs of chromosomes
• 3 Gbps (bps base pairs) , only 3 are genes
• 25,000 genes
• Yeast
• 16 chromosomes
• 20 Mbps
• 6,500 genes

9
Regulation of Expression

• Each cell contains an identical copy of the whole
  genome - but utilizes only a subset of the genes
  to perform diverse, unique tasks
• Most genes are highly regulated
• their expression is limited to specific tissues,
  developmental stages, physiological condition
• Main regulatory mechanism transcriptional
  regulation

10
Transcriptional regulation

• Transcription is regulated primarily by
  transcription factors (TFs) proteins that bind
  to DNA subsequences, called binding sites (BSs)
• TFBSs are located mainly (not always!) in the
  genes promoter the DNA sequence upstream the
  genes transcription start site (TSS)
• BSs of a particular TF share a common pattern, or
  motif
• Some TFs operate together TF modules

TSS
11
TFBS motif models

• Consensus (degenerate) string

AC
CG
ACT
T
gene 1
gene 2
AACTGT
gene 3
CACTGT
gene 4
CACTCT
gene 5
CACTGT
gene 6
gene 7
gene 8
gene 9
AACTGT
gene 10

• Statistical models
• Motif logo representation

12
Human G2M cell-cycle genesThe CHR NF-Y module
CDCA3 (trigger of mitotic entry
1) CTCAGCCAATAGGGTCAGGGCAGGGGGCGTGGCGGGAAGTTTGAAAC
T -18 CDCA8 (cell division cycle associated
8) TTGTGATTGGATGTTGTGGGA25bpTGACTGTGGAGTTTGAAT
TGG 23 CDC2 (cell division control protein 2
homolog) CTCTGATTGGCTGCTTTGAAAGTCTACGGGCTACCCGATTG
GTGAATCCGGGGCCCTTTAGCGCGGTGAGTTTGAAACTGCT
0 CDC42EP4 (cdc42 effector protein
4) GCTTTCAGTTTGAACCGAGGA25bpCGACGGCCATTGGCTGCT
GC -110 CCNB1 (G2/mitotic-specific cyclin
B1) AGCCGCCAATGGGAAGGGAG30bpAGCAGTGCGGGGTTTAAA
TCT 45 CCNB2 (G2/mitotic-specific cyclin
B2) TTCAGCCAATGAGAGT15bpGTGTTGGCCAATGAGAAC15
bpGGGCCGCCCAATGGGGCGCAAGCGACGCGGTATTTGAATCCTGGA
10 BSs are short, non-specific, hiding in
both strands and at various locations along the
promoters
TFs NF-Y , CHR
13
Protein Binding MicroarraysBerger et al, Nat.
Biotech 2006

• Generate an array of double-stranded DNA with all
  possible k-mers
• Detect TF binding to specific k-mers

13
14
PBM (2)
14
15
PBM - implementation

• Use 60-mers (Agilent) 25nt constant primer
  35nt variable region
• De Bruijn seq of all 10-mers (410 long) split
  into 35nt long fragments with 9nt overlap
• 40K probes
• For each 8-mer, combine signals from all probes
  that contain it (or differ in 1nt) to obtain its
  binding score

15
16
The computational challenge

• Input PBM data (sequences and binding scores) of
  one TF.
• Goal Find a motif (PWM) that is the binding site
  of that TF.
• Intuition sequences that match the motif (on one
  of the two possible strands!) are expected to
  have high binding scores.

17
2. The project
18
General goals

• Research
• - Learn about known solutions
• - Trial and error with training data
• Develop software from A-Z
• Design
• Implementation (Optimization)
• Execution analysis of test data
• A taste of bioinformatics
• Have fun
• Get credit

19
The computational task

• Given a set of PBM data of different TFs.
• Find the binding site motif in PWM format of each
  TF.
• Main challenges
• Performance (time, memory)
• Accuracy

20
Input

• File with 41,923 lines, each containing a probe
  sequence of length 35 and binding intensity.
• ltsequence 35bpgt \t ltintensitygt \n

21
Input (II)

• For the training data, an additional PWM file
  will be supplied for each PBM data set.
• A ltfreq1gt ltfreq2gt ltfreq10gt
• C ltfreq1gt ltfreq10gt
• G
• T
• Separated by \t and \n.
• All lines must contain same number of frequencies
  (10 is just an example).

22
Input (III)

• You will be given
• 10 training sets (PBM data PWM)
• 4 test sets (PBM data). You have to provide the
  PWM.
• In the final project presentation, you will be
  given an online test set (PBM data) and your
  software will be applied to it.

23
Output

• A PWM file describing the binding site found in
  the given PBM file.
• The PWM in motif logo format (i.e. displayed on
  the screen).
• The file logo.zip contains a java package with
  the code that will easily display your motif.

bits 2 - entropy
24
Output (II)

• Show graphically how well your motif predicts
  the binding intensity.
• One example (note its not PWM)

25
Ranking 8-mers

• One possible way to start rank the 8-mers in
  some way. Scores for example
• 1. Signal average.
• 2. Signal median.
• You can think of other scores that incorporate
  more information, e.g. position in probe
  sequence.
• This is just an example. You can think of other
  ways to start.

26
Alignment procedure

• Then, you can align the significant 8-mers.
• You may take into account the relative score.
• Dont forget about the reverse complement!
• Example Cebpb TF

27
Enrichment scores

• To test how good your motif is, you can use an
  enrichment score.
• An enrichment score tests how good the motif
  distinguishes between high-ranking probes and the
  rest of the probes.

28
Hypergeometric probability
total not drawn drawn
m m - k k white
N - m N k - n - m n - k black
N N - n n total
29
Hypergeometric enrichment score

• Let B and T (T B) denote the BG and target
  sets, respectively, and let b and t denote the
  subset of probes from the BG and target set,
  respectively, that contain at least one
  occurrence of the motif.

30
Hypergeometric score (2)

• The HG enrichment score computes the probability
  of observing at least t target sequences with a
  motif occurrence, under the null hypothesis that
  the probes in the target set were drawn randomly,
  independently, and without replacement from the
  BG set.
• Code is provided in math.zip

31
Wilcoxon-Mann-Whitney (WMW) enrichment score

• Foreground probes are all those containing a
  match, background are all the others.
• B and F are the sizes of background and
  foreground, respectively.
• ?B and ?F are the sums of the background and
  foreground ranks.
• Read more in supplementary info (Berger06).

32
Deciding the length of the motif

• Another challenge is to decide the length of the
  motif.
• Most binding site are 6-12 bp long.
• You should consider the information each position
  contains and decide on the length accordingly.

33
Scoring your PWM

• One way to score your motif is by ranking the
  probe sequences according to a match score.
• You may use the given code for match score.
• Compare the ranking of the probes you got to the
  ranking according to binding intensities. There
  are different correlation score for that.

34
Match Score between PWMs

• Already implemented for you
• Euclidian Distance
• Pearson Correlation Coefficient
• KL Divergence

35
Implementation

• Java (Eclipse) Linux (Other languages are
  possible, but will not participate in bonus).
• Input one single argument PBM filename
• Output PWM file, PWM presented in logo and
  graphical presentation of PWM matching
  distribution among probes.
• Packages for motif logo and statistical scores
  will be supplied
• Time performance will be measured
• Reasonable documentation
• Separate packages for data-structures, scores,
  GUI, I/O, etc.

36
Submission

• Printed design document.
• Printed code for comments and remarks.
• Printed results document for each test set PWM
  logo how good your result in terms of
  correlation to the probes ranks.
• 4 PWM files, e.g. Test_1.pwm (submitted by
  email).
• Executable for the online test.

37
Grade

• 20 for the design
• 30 for the implementation (20 for modularity,
  clarity, documentation, 10 for efficiency)
• 30 for the performance and experimental results
  (20 for the accuracy on the 4 test queries and
  10 for the accuracy on the online test query)
• 20 for the final report and presentation
• 10 bonus to the group with the most accurate
  results
• 10 bonus for the group with the fastest
  implementation

38
Bonus grading

• Accuracy will be determined using the provided
  code that compares two PWMs.
• We will take the average of runs on several
  different PBM data sets.
• Running time will be measured in java
  implementation, and the average will be taken.

39
Schedule

• First progress report 23/11
• Design document 21/12
• Final presentation 16/2
• We shall meet with each group on each of these
  dates mark your calendars!
• Schedule can be made earlier if you are ready.
• You are always welcome to meet us. Contact us by
  email.

40
Design document

• Due in week 12 (21/12).
• 3-5 pages (Word), Hebrew/English
• Briefly describe main goal, input and output of
  program
• Describe main data structures, algorithms, and
  scores.
• Meet with me before submission.

41
Reference

• Berger MF, Philippakis AA, Quershi AM, He FS,
  EstepIII PW, Bulyk ML. Compact, universal DNA
  microarrays to comprehensively determine
  transcription-factor binding site specificities.
  Nature biotechnology. 20063381429-1435.
• Very important! Read the_brain.bwh.harvard.edu/UP
  BMseqn/suppl_methods.doc
• Chen X, Hughes TR, Morris Q. RankMotif a
  motif-search algorithm that accounts for relative
  ranks of K-mers in binding transcription factors.
  Bioinformatics. 2007 Jul 123(13)i72-79.

42
Fin