testGeneFamilies.binomial {ShotgunFunctionalizeR} | R Documentation |
Test for significant gene families using the binomial model.
testGeneFamilies.binomial(data)
data |
$2times n$ matrix containing counts |
testGeneFamilies.binomial
performs an exact binomial test for each row
in the data matrix. Assume that $X_1sim Po(M_1p_1)$ and $X_2sim
Po(M_2p_2)$ where $M_1$ and $M_2$ are the total number of
observations from two metagenomes. It follows that
[
X_1|X_1+X_2=n sim Bin(n,frac{M_1p_1}{M_1p_1+M_2p_2})
]
and the binomial test is testing whether $p_1=p_2$. The test can be seen as an approximation of the
hypergeometric counterpart (testGeneFamilies.hypergeometric
).
testGeneFamilies.binomial
is usually called
implicitly by using the 'binomial' method in testGeneFamilies
.
testGeneFamilies.binomial
returns a list with the following
elements
PValues |
A list of p-values |
Rankings.detail |
A list additional elements containing details about the ranking |
testGeneFamilies
,
testGeneFamilies.hypergeometric
,
testGeneFamilies.gaussian
,
testGeneFamilies.poisson
counts<-cbind(rpois(10, 100), rpois(10, seq(50,300, length=10))) testGeneFamilies.binomial(counts)