Enrichment test in cancer subtypes

	Invasive	Proliferative	Metabolic
Hypermethylated	4110	1536	988
Hypomethylated	625	1263	284

Q1: Is the Invasive subtype enriched with Hypermethylated CpGs?

q <- 4110 # number of success (white balls) drawn

m <- 4110+1536+988 # number of success (white) in the urn

n <- 625+1263+284 # number of fail (black) in the urn

k <- 4110+625 # number of balls drawn out

p.value <- phyper(q-1,m,n,k, lower.tail=F)

# why q-1, because we want to have Pr(X>=x) instead of Pr(X>x)

# ref: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2724271/ , search "phyper"

p.value

= 2.975069e-162

YES

Conclusion: The Invasive subtype is significantly enriched with Hypermethylated CpGs.

Q2: Is the Invasive subtype enriched with Hypomethylated CpGs?

q <- 625 # number of success (white balls) drawn

m <- 625+1263+284 # number of fail (black) in the urn

n <- 4110+1536+988 # number of success (white) in the urn

k <- 4110+625 # number of balls drawn out

p.value <- phyper(q-1,m,n,k, lower.tail=F)

p.value

=1

No

Q3: Is the Proliferative subtype enriched with Hypermethylated CpGs?

q <- 1536

m <- 4110+1536+988

n <- 625+1263+284

k <- 1263+1536

p.value <- phyper(q-1,m,n,k, lower.tail=F)

p.value

 = 1

No

Q4: Is the Proliferative subtype enriched with Hypomethylated CpGs?

q <- 1263

m <- 625+1263+284

n <- 4110+1536+988

k <- 1263+1536

p.value <- phyper(q-1,m,n,k, lower.tail=F)

p.value

= 3.765696e-193

YES

Conclusion: The Proliferative subtype is significantly enriched with Hypomethylated CpGs.

Q5: Is the Metabolic subtype enriched with Hypermethylated CpGs?

q <- 988 # number of success (white balls) drawn

m <- 4110+1536+988 # number of success (white) in the urn

n <- 625+1263+284 # number of fail (black) in the urn

k <- 988+284 # number of balls drawn out

p.value <- phyper(q-1,m,n,k, lower.tail=F)

p.value

= 0.01919936

YES

Conclusion: The Metabolic subtype is significantly enriched with Hypermethylated CpGs.

Q6: Is the Metabolic subtype enriched with Hypomethylated CpGs?

q <- 284 #

m <- 625+1263+284

n <- 4110+1536+988

k <- 988+284

p.value <- phyper(q-1,m,n,k, lower.tail=F)

p.value

= 0.9839128

No

See also:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2724271/
An integrative genomics approach identifies Hypoxia Inducible Factor-1 (HIF-1)-target genes that form the core response to hypoxia