bda (version 13.4.13)

fit.mlnorm: Fit a log-normal distribution to (binned) NGS data

Description

Fit a log-normal distribution to (binned) NGS data

Usage

fit.mlnorm(x, method="mixed", k=1, gridsize=1000,optim=TRUE)
  fit.lognormal(x, method='mle')

Arguments

x

Next-Generation Sequencing data or similar. Could have a lot of zeros but negative values are not allowed

method

choose fitting method. For fit.mlnorm, options include "Monte Carlo method", and "Maximum Likelihood method". For fit.lognormal, two options: 'mle' or 'percentile' -- percentile matching method.

gridsize

size of grid where densities will be evaluated.

k

Number of components.

optim

Choose the best model if 'true'.

Details

We fit a mixture model to take care the zero values using component-0.

References

AS 254, ...

Examples

# NOT RUN {
 # To be updated.
 x = rexp(100,1)
 x = c(rep(0,20),x)
 out = fit.NGS(x)
 plot(out)
# }