ProPhET

ProPhET is a probabilistic programming tool written in Julia and Turing that can be used to estimate the measure of a phylogenetic signal. The ispiration for the tool came form the work our group did on the inference of SARS-CoV-2 sublineage variants identification, VERSO: A comprehensive framework for the inference of robust phylogenies and the quantification of intra-host genomic diversity of viral samples (cfr., the VERSO repository). The original Bayesian model used in ProPhET is described in Measuring phylogenetic signal between categorical traits and phylogenies.

Input

ProPhET takes in input the following files:

• inference: the VERSO step one output
• clonal_variants: the VERSO input
• names: exported names of rows and columns of the following objects:
  • B matrix
  • C matrix
  • corrected_genotypes
  • clonal_variants

Input preparation

To build names it is possible to export the data from R:

names = list("B_rows" = rownames(inference$B),
             "B_cols" = colnames(inference$B),
             "C_rows" = rownames(inference$C$Experiment_1),
             "G_rows" = rownames(inference$corrected_genotypes),
             "G_cols" = colnames(inference$corrected_genotypes),
             "CL_rows" = rownames(clonal_variants),
             "CL_cols" = colnames(clonal_variants))
save(names, file="~/names.RData")

Running ProPhET

ProPheT can be run form the CLI as follows:

$ julia ProPhET.jl inference.RData clonal_variants.RData names.RData

Is also possible to specify the following optional parameters:

Parameter	Effect	Type	Default
--sampler	Set sampler	String (NUTS or MH)	NUTS
--n	Set the number of generated samples	Integer	40000
--lambda	Set the lambda hyperparameter	Float64	0.1
--error	Set the error rate for mutations selection	Float64	0.030
--pltName	Set the plots filename prefix	String	plot_

Inference

ProPhET computes the entropy value of each node and uses the entropy vector to perform a Bayesian inference. The sampling process can be built with the following methods:

--sampler value	Sampler
MH	Gibbs(MH(:alpha), MH(:beta))
NUTS	NUTS()

Output

The output of ProPhET is a real number $\delta$. An high value of $\delta$ means that the phylogenetic signals are more informative. A low value of $\delta$ could suggest that a case of homoplasia could be present in the phylogenetic tree.