Predicting effect of SNPs and de novo mutations on splicing

1
Predicting effect of SNPs and de novo mutations
on splicing

• presented by
• Alexander Tchourbanov
• Biology Department
• New Mexico State University

2
Motivation

• Recently, high throughput genotyping methods
  became available
• High-density 500K chips are available for
  genotyping (Illumina Hap550, Affymetrix 5.0)
• Genome resequencing (SOLID Applied Biosystems,
  Solexa/Illumina genome analyzer, Roche 454 FLX)
• Researchers, interested to understand genetic
  risk factors contributing to a disorder,
  routinely genotype patients

3
Motivation

• Many SNPs have been associated with
  predisposition to various diseases (Breast
  cancer, Alzheimer's, Multiple sclerosis, etc.)
• Only fraction of actual SNPs are genotyped with
  chips
• Some SNPs with significantly low P-values have
  been associated through LD with affected
  haplotypes
• Fraction of associated SNPs are causal variants
• There is a growing evidence that Autism Spectrum
  Disorder (ASD) could be triggered by de novo
  mutations absent in both parents

4
Types of SNPs

• Several classes of variants to consider
• Single Nucleotide Polymorphisms (SNPs)
• Deletion/Insertion polymorphisms (DIPs)
• Simple Tandem Repeat polymorphisms (STRs)
• Named polymorphisms (e.g., Alu/ dimorphisms)
• Multinucleotide polymorphisms (MNPs)

5
SNPs distribution

• 6 million SNPs are located in human gene loci
  (dbSNP build 129)
• 63 intronic
• 11 untranslated region
• 1 nonsynonymous
• 1 synonymous
• 24 ?2 kBp from a gene
• lt1 splice site
• lt1 unknown coding variant

6
What are the common disease causing variants?

• SNPs are defined as former mutations with gt1 of
  population penetrance
• According to Human Gene Mutation Database HGMD
  (http//www.hgmd.cf.ac.uk)
• 49,806 mutations are missence/nonsense
• 8,548 mutations have consequences in mRNA
  splicing
• Many missence/nonsence mutations are eliminated
  by purifying selection and never make it to SNPs

7
Splicing components
Image credit Understanding alternative splicing
towards a cellular code Arianne J. Matlin,
Francis Clark and Christopher W. J. Smith, Nature
Reviews Molecular Cell Biology 6, 386-398 (May
2005)
8
Existing elements
9
Orthologos blocks from UCSC GB

• 2,333,379 extended exons from 23 Tetrapoda
  organisms were obtained
• A number of experimental reports showed that
  genes from distantly related Tetrapoda organisms
  were correctly expressed and post-transcriptionall
  y modified in transgenic animals (Capetanaki Y et
  al. Proc Natl Acad Sci USA 1989, Jacobs GH et
  al. Science 2007)
• The genes encoding well-known RNA binding
  proteins involved in splicing regulation are
  enriched with ultraconserved elements (Bejerano
  G. et al.Science 2004)

10
Counting oligos
11
Comparing oligo counts
12
Elements found

• Using the orthologous exons available for 23
  Tetrapoda organisms we have identified 2,546
  unique splicing regulatory elements.
• Among these elements 203 (7.97) 3SS and 177
  (6.95) 5SS supporting motifs are novel and have
  not been previously reported in systematic
  screens detecting such elements.
• Among our predicted elements, 41.08 of sequences
  were heptamers and 51.81 were octamers and only
  6.76 hexamers and 0.35 pentamers

13
Example of 5SS ISEs found
14
Example of LOD profiles (5SS ISE)
15
Optimal exon length

• Depends on flanking 5SS and 3SS strengths

16
5GC SS Bayesian sensor
17
Exon scoring method

• LOD scores associated with 5SS,3SS, exonic
  length, competing SSs and Enhancer/Silencer
  signals are combined towards an exon strength

18
IVS22delC mutation

• gtIVS22delC
• ttcggataagacaaagattttatataatattttgaaaacattaaat
  aatt tgtcattcctttatttcctttattttagCTTCGCAGAATCAAGAA
  CGGCTATGTGCGTTTAAAGATCCGTATCAGCAAGACCTTGGGATAG/GTG
  AGAGTAGAATCTCTCATGAAAATGGGACAATATTATGCTCGAAAG/GTAG
  CACCTGCTATGGCCTTTGGGAGAAATCAAAAGGGGACATAAATCTTGTAA
  AACAAGg(c)aagtgatactttccttacctgaaatgactgtgttttatac
  aattgatatttatctaaaaaggacatgggagtatgttaaaatcctgttca
  gaaaaacagtgaatttaaaagtgtatatataaagccaggtgtggtggctc
  atgcctgtaattccagcacttttcgaggctgaggtgggcggatcacttga
  ggccaggagtttgagaccagcctgggtaataacatggtgaaaccccgt

19
Example of SpliceScan II predicting effects of
mutations

• An example of successfully predicted effect of
  mutation IVS22delC causing familial pulmonary
  arterial hypertension (Cogan JD et al Am J
  Respir Crit Care Med 2006)
• Another example of SpliceScan II correctly
  predicting the effect of IVS10-6del34 micro
  deletion causing gastrointestinal stromal tumors
  (Chen LL et alOncogene 2005 )

20
SpliceScan II performance on mutations
21
Effect of rs849563 (Autism associated SNP)

• There is a change in annotated exon potential
  here
• rs849563 changes the exon sharing one boundary
  with annotated exon gi41872561refNM_201266.1
  2433-2577 where the exon score changes 0.60-gt0.19

22
Effect of rs885747 (Autism associated SNP)

• There is a change in annotated exon potential
  here
• rs885747 changes the exon sharing one boundary
  with annotated exon gi194097340refNM_002616.2
  1627-1735 where the exon score changes
  0.30-gt0.49

23
SNPs performance
24
SpliceScan II tool

• SpliceScan II tool http//splicescan2.lumc.edu/
• Is more sensitive than existing splicing
  simulators (NetUTR, ExonScan)
• Uses novel 5 GC SS Bayesian sensor
• Method allows predicting aberrant splicing events
  associated with genomic variants
• ACGMAP companion database http//www.stritch.luc.e
  du/node/375

25

• Thanks!