Aspects of Genetics and Genomics in Cancer Research

1
Aspects of Genetics and Genomics in Cancer
Research

• Li Hsu
• Biostatistics and Biomathematics Program
• Fred Hutchinson Cancer Research Center

2
Outline

• Cancer facts
• Linkage analysis of family studies
• Genome-wide association studies

3
(No Transcript)
4
Etiology of Cancer

• The etiology of cancer is multifactorial, with
  genetic, environmental, medical, and lifestyle
  factors interacting to produce a given
  malignancy.
• The breakthroughs in high throughput genotyping
  technologies have made it possible for
  systematically identifying genes that are
  responsible for disease occurrence.

5
BRCA1 and Breast Cancer

• BRCA1 (breast cancer 1) is a human gene that
  belongs to a class of genes known as tumor
  suppressors, which maintains genomic integrity to
  prevent uncontrolled proliferation. Variations in
  the gene have been implicated in a number of
  hereditary cancers, namely breast, ovarian and
  prostate. The BRCA1 gene is located on the long
  (q) arm of chromosome 17 at 38Mb.

6
Probability of developing breast cancer by age
(Chen et al. 2009)
carriers
Non-carriers
7
Probability of Developing Breast Cancer for BRCA1
carriers
Average Person BRCA1 Carrier
Age 50 2.1(1.7-2.7) 18.8(8.2-2.3)
Age 60 4.1(3.4-5.0) 31.3(14.3-61.2)
Age 70 7.2(6.0-9.0) 45.4(22.7-74.3)
Age 80 10.2(8.4-12.5) 54.9(30.4-81.4)
8

• How was BRCA1 found?

9
(No Transcript)
10
Linkage Analysis
11

• Assume disease gene (D) is rare with full
  penetrance

12
Linkage Analysis (continued)

• Disease allele (D) originally in chromosome with
  allele 3
• How often does D co-segregate with allele 3
  (non-recombinant)?

13

• Assume disease gene (D) is rare with full
  penetrance

14
Linkage Analysis (continued)

• Disease allele (D) originally in chromosome with
  allele 3
• How often does D co-segregate with allele 3
  (non-recombinant)?
• 5 meiosises
• How often is D separated from allele 3
  (recombinant)?

15

• Assume disease gene (D) is rare with full
  penetrance

16
Linkage Analysis (continued)

• Disease allele (D) originally in chromosome with
  allele 3
• How often does D co-segregate with allele 3
  (non-recombinant)?
• 5 meiosises
• How often is D separated from allele 3
  (recombinant)?
• 1 meiosis

17
Likelihood function

• Set a parameter ? which measures the distance
  between allele 3 and D by how frequently they
  recombine.
• The likelihood function L(?) (1- ?)5 ?
• The maximum likelihood estimate is 1/6
• LOD log10 L(1/6)/L(1/2)
• 0.63
• LOD for 7 families 7x0.63 4.41

18
Issues

• Linkage analysis has narrowed down to a region
  about 1Mb. However it took another four years
  before the BRCA1 gene was mapped.
• Reduced penetrance, phenocopy, and genetic
  heterogeneity are among the factors that limit
  the success of the linkage analysis.
• Relevance of the findings to the population at
  large.

19
Genome-Wide Association Studies(GWAS)

• The Human Genome Project began in 1990 and
  completed in 2003.

20
Part of sequence from Chromosome 7

• AGACGGAGTTTCACTCTTGTTGCCAACCTGGAGTGCAGTGGCGTGATCTC
  AGCTCACTGCACACTCCGCTTTCC/TGG
• TTTCAAGCGATTCTCCTGCCTCAGCCTCCTGAGTAGCTGGGACTACAGTC
  ACACACCACCACGCCCGGCTAATTTTTG
• TATTTTTAGTAGAGTTGGGGTTTCACCATGTTGGCCAGACTGGTCTCGAA
  CTCCTGACCTTGTGATCCGCCAGCCTCT
• GCCTCCCAAAGAGCTGGGATTACAGGCGTGAGCCACCGCGCTCGGCCCTT
  TGCATCAATTTCTACAGCTTGTTTTCTT
• TGCCTGGACTTTACAAGTCTTACCTTGTTCTGCCTTCAGATATTTGTGTG
  GTCTCATTCTGGTGTGCCAGTAGCTAAAA
• ATCCATGATTTGCTCTCATCCCACTCCTGTTGTTCATCTCCTCTTATCTG
  GGGTCACA/CTATCTCTTCGTGATTGCATTC
• TGATCCCCAGTACTTAGCATGTGCGTAACAACTCTGCCTCTGCTTTCCCA
  GGCTGTTGATGGGGTGCTGTTCATGCCT
• CAGAAAAATGCATTGTAAGTTAAATTATTAAAGATTTTAAATATAGGAAA
  AAAGTAAGCAAACATAAGGAACAAAAAG
• GAAAGAACATGTATTCTAATCCATTATTTATTATACAATTAAGAAATTTG
  GAAACTTTAGATTACACTGCTTTTAGAGAT
• GGAGATGTAGTAAGTCTTTTACTCTTTACAAAATACATGTGTTAGCAATT
  TTGGGAAGAATAGTAACTCACCCGAACA
• GTGTAATGTGAATATGTCACTTACTAGAGGAAAGAAGGCACTTGAAAAAC
  ATCTCTAAACCGTATAAAAACAATTACA
• TCATAATGATGAAAACCCAAGGAATTTTTTTAGAAAACATTACCAGGGCT
  AATAACAAAGTAGAGCCACATGTCATTT
• ATCTTCCCTTTGTGTCTGTGTGAGAATTCTAGAGTTATATTTGTACATAG
  CATGGAAAAATGAGAGGCTAGTTTATCAA
• CTAGTTCATTTTTAAAAGTCTAACACATCCTAGGTATAGGTGAACTGTCC
  TCCTGCCAATGTATTGCACATTTGTGCCC
• AGATCCAGCATAGGGTATGTTTGCCATTTACAAACGTTTATGTCTTAAGA
  GAGGAAATATGAAGAGCAAAACAGTGCA
• TGCTGGAGAGAGAAAGCTGATACAAATATAAATGAAACAATAATTGGAAA
  AATTGAGAAACTACTCATTTTCTAAATT
• ACTCATGTATTTTCCTAGAATTTAAGTCTTTTAATTTTTGATAAATCCCA
  ATGTGAGACAAGATAAGTATTAGTGATGGT
• ATGAGTAATTAATATCTGTTATATAATATTCATTTTCATAGTGGAAGAAA
  TAAAATAAAGGTTGTGATGATTGTTGATTA

21
Genome-Wide Association Study

• 550,000 SNPs on an array
• 2000 diseased individuals (colon cancer cases)
  and 2000 normal individuals
• Genotype all DNAs for 550,000 SNPs
• That is 2 billion genotyping!

22
GWAS on Type 2 Diabetes (Steinthorsdottir et al.,
2007, Nature Genetics)
Cases Controls
AA 809 3049 3858
Aa 509 1917 2426
aa 81 305 385
1398 5271 6669
Cases Controls
AA 751 3107 3858
Aa 539 1887 2426
aa 108 277 385
1398 5271 6669

• Expected count for cases if AA is not associated
  with the disease. First, calculate the frequency
  of AA genotype in both cases and controls
  combined
• freq 3858/6669 57.85
• For 1398 cases, we expect to see 139857.85809
  individuals having genotype AA.

23
GWAS on Type 2 Diabetes

• The chi-square statistic is calculated by finding
  the difference between each observed and expected
  for each cell, squaring them, dividing each by
  the expected, and taking the sum of the results.
• (757-809)2/809(3107-3049)2/3049
• Compare the value to a standard chi-square
  distribution with degrees of freedom (
  rows-1)( col -1) 2.
• The p-value for this SNP is 6.772e-5.

24
Issues

• Too many SNPs!
• Identifying gene-gene and gene-environmental
  interactions are now possible.

25
Germline mutations account for only a small
portion of cancer cases.
http//envirocancer.cornell.edu/FactSheet/General/
fs48.inheritance.cfm
26
Summary

• The amount of the data that have been generated
  increases exponentially in the last few years.
• This creates a great demand on efficient and
  valid computational and statistical methods and
  tools for picking the needles from a haystack.