Q: The function Q for the EM algorithm for the negative binomial...

In johannabertl/SelectionMix: Negative binomial mixture model

Description

The function negQ = -Q can be used for minimization.

Usage

Q(theta, theta.prime, x.syn, x.non, cvec)

Arguments

theta	c(alpha, beta, p1, p2, ..., pk-1)
theta.prime	c(alpha.prime, beta.prime, p1.prime, p2.prime, ..., p(k-1).prime)
x.syn	vector of synonymous mutation count
x.non	vector of non-synonymous mutation count
cvec	vector of k positive values c(c1, ..., ck)

Author(s)

Johanna Bertl

Examples

### Example with 2 components ###

# Simulate dataset of synonymous and non-synonymous mutations

c1 = 0.5; c2 = 10
p1 = 0.2; p2 = 0.8
alpha = 10
beta = 5

mutations = rnegbinmix_syn_nonsyn(n=3000, alpha=alpha, beta=beta, c=c(c1, c2), p = c(p1, p2))

# Q

Qtest = Q(theta = c(5, 1, 0.5), theta.prime = c(5, 1, 0.5), x.syn = mutations$Syn, x.non = mutations$Non, c = c(c1, c2))


### Example with 3 components ###

c1 = 0.01; c2 = 1; c3 = 100
p1 = 0.33; p2 = 0.33; p3 = 0.34
alpha = 10
beta = 5

mutations = rnegbinmix_syn_nonsyn(n=3000, alpha=alpha, beta=beta, c=c(c1, c2, c3), p = c(p1, p2, p3))

# Q

Qtest = Q(theta = c(5, 1, 0.3, 0.3), theta.prime = c(5, 1, 0.3, 0.3), x.syn = mutations$Syn, x.non = mutations$Non, c = c(c1, c2, c3))

johannabertl/SelectionMix documentation built on May 3, 2019, 4:03 p.m.