Transcript Whole Genome Polymorphism Analysis of Regulatory Elements in

Whole Genome Polymorphism
Analysis of Regulatory Elements in
Breast Cancer
Jacob Biesinger
Dr. Garry Larson
City of Hope
AAGTCGGTGATGATTGGGACTGCTCT[C/T]AACACAAGCGAGATGAAGAAACTGA
Topics Covered Today




Molecular Cause of Genetic Disease
Cancer and Gene Regulation
Combining Data: Bioinformatics
Progress So Far

http://medicine.osu.edu/lend/Portfolios/0506/AR Port/files/SICKLE CELL WEBSITE/whatissickle.htm
Single Nucleotide Polymorphisms
and Genetic Disease

SNPs in coding regions:
Phe
Pro
Glu
Val
Thr STOP
Ser
ATGCCGGCTTACCATA A
T TCTACCTAAATCCGGT

Genetic disease may also be caused by
differential expression of vital proteins
Promoter Binding
Mechanism
Sickle Cell Anemia
TGTAGA
ATGCCGGCTTACCATA T
ATCTACCTAAATCCGGT
Micro RNA Binding
Mechanism
Protein Coding Region
Chunky sheep from miRNA
binding site destruction
Untranslated region
Nature Rev. Genet. 5, 202–212
(2004)
Breast Cancer Expression
Normal Breast
Expression


Breast Tumor
Expression
Tumor expression patterns are
extremely divergent from
normal cells
Could SNPs in regulatory
regions of genes associated
with breast cancer explain their
overexpression in tumors?
http://genome-www.stanford.edu/breast_cancer/cell_line_review2001/images/figure2.html




Expression patterns in cancers
gives two categories: Estrogen
Receptor + and ERRecent metaanalysis pooled
tumor expression data for 9
studies and >15,000 genes
Top 1% ER+ > ER- 150 genes
Top 1% ER+ < ER- 150 genes
Consistency across studies
Statistical Search for Dysregulated
Genes
Normalized expression difference
between ER+ and ER-
Regulation Motifs


Which TF binding sites exist in our selected genes?
A recent study identified motifs conserved in
regulatory regions across 4 organisms
lymphocyte transmembrane adaptor 1
Promoter motifs:


123 known motifs
174 phylogenetically
conserved
Downstream motifs:
 273 conserved 3’ UTR
 343 conserved miRNA 6mer
 368 conserved miRNA 7mer
Motif Search

Use Python and UCSC Genome Browser to:




Get promoter region DNA (2kb upstream from
transcription start site (TSS) + max of 2kb downstream
of TSS, limited by translation start)
Get 3’ untranslated region RNA
Search for motifs on + and – strand
Results for Top 1% up and down:


22206 known motif hits
23475 phylo motif hits



9559 3’ UTR hits
42846 6mer hits
11719 7mer hits
SNP Databases

HapMap
~4 million
CGEMS
~550k
SNP information is coming from two databases:


HapMap- Four groups (270 total people) genotyped for
same SNPs
CGEMS- Breast Cancer association study, complete
with p-values. A late-comer to our study (June 2007)
Mapping SNPs
Gene Promoters and 3’ UTR
HapMap
~4 million
CGEMS
~550k
Motif Matches

Use MSSQL 2003 and Python (pymssql) to
perform a join of dbSNP, HapMap and
CGEMS SNPs with regulatory motifs
Verify Motif Significance

How do we know that these motifs are significant?
Hypothesis: Due to negative selection, there will be
fewer SNPs in motifs than in random areas within
the same region.
Method: Contrast how many motifs have at least one
SNP in them against how many of 100 random
sequences from the same region have at least one
SNP in them
Motif Counting Results
Known Top 1%
Actual
Random
1-Sided PMotif with Snp
Motif without Snp Total
Value
97
18394
18491 0.000009494
14630
1834470
1849100
Total
14727
1852864
1867591
1-Sided PValue
Phylo Top 1%
Actual
130
19363
19493
Random
16499
1913438
1929937
Total
16629
1932801
1979430


0.001889
3’ UTR results not yet available
There is a significant difference between motifs
and random sequences.
CGEMS Results

A number of SNPs that fall within motifs are associated
with Breast Cancer
Highest ranking was 1514 out of 550,000

Further analysis required to say if significant

Thanks!







SoCalBSI mentors
City of Hope
Dr. Garry Larson
Dr. David Smith
Dr. Päl Sætrom
Cathryn Lundberg
All the SoCalBSI students!
Funded by: