dist_prob: Fit a distribution to pair_wise bottlenecks

In jtmccr1/HIVEr: Analysis of transmission and intrahost iSNV data taken as part of the HIVE cohort

Description

This funciton returns the negative log likelihood for a distribution of bottlenecks given a data frame of fits on a per-pair basis. This can be used with mle from the bbmle package.

Usage

dist_prob(data, weight, ddist, ...)

Arguments

data	A tibble or data frame as outputed from trans_fit
weight	A tibble or data frame providing the weighting factors for each pair. It should have columns pair Id and weight_factor
ddist	The pdf or pmf function for the distribution to test.
...	The parameters to pass to the ddist function

Value

the negative Log likelihood weighted by the weighted data frame

Examples

require(magrittr)
require(dplyr)
fit <-trans_fit(small_trans,100,"PA",threshold = NULL,acc=NULL,pair_id)
weights<-small_trans %>% group_by(pair_id) %>%
summarize(donor_mutants = length(which(freq1>0 & freq1<0.5))) %>%
 mutate(weight_factor = max(donor_mutants)/donor_mutants)
dist_prob(fit,weights,dzpois,lambda=1)

jtmccr1/HIVEr documentation built on May 29, 2019, 1:50 a.m.