phi.ml: Estimates dispersion parameter phi
In DavidKLim/FSCseq: Feature Selection and Clustering of RNA-seq Count Data

Uses theta.ml2 and theta.mm functions to tryCatch any unstable estimations

1	phi.ml(y, mu, dfr, weights, t0, limit, trace)

`y`	vector: gene read counts for one fixed gene
`mu`	vector: mean. must be same length as y
`dfr`	numeric: degrees of freedom for theta.mm()
`weights`	vector: must be same length as y
`limit`	numeric: max number of iterations for theta.ml() and theta.mm()
`trace`	logical: True/False to show trace output from theta.ml