Dataset: NWHI_HABC_Simulations
Deployment: Chan_model

hBayeSSC download available here: https://github.com/UH-Bioinformatics/hBayeSSC

hBayeSSC is a Python script that wraps around Serial SimCoal in order to simulate a multi-taxa community undergoing a coordinated demographic expansion.

Requirements:

The applications required to produce a set of simulations with multitaxa summary statistics for the hABC analysis described in Chan et al. 2014:

- BayeSSC - Serial Simcoal
- Python 2.x >= 2.4
- hBayeSSC.py
- msReject

Input files

The input files needed to produce a set of simulations with multitaxa summary statistics for the hABC analysis described in Chan et al. 2014:

- A table of observed summary statistics for each taxon in the community. (details)
- An input par file for serial simcoal. (details)

Observation summary statistics:

Sample observation file:

The table of observed summary statistics consists of columns with the following header names. hBayeSSC replaces the appropriate line in the par file with these values:
Column name  =  Description
species = Name of taxa
nsam = number of samples to be simulated
nsites = Number of base pairs
tstv = % transitions
gamma = Gamma shape parameter
gen = Numbers of years per generation
locuslow = Low estimate of the locus mutation rate per generation
locushigh = High estimate of the locus mutation rate per generation
Nelow = Low estimate for effective population size
Nehigh = High estimate for effective population size
SegSites = Segregating sites
nucdiv = Nucleotide diversity
Haptypes = Number of haplotypes
HapDiver = Haplotypic diversity
TajimasD = Tajima's D
F* = Fu's F
A more complete description of these values can be found on the BayeSSC website.

par file

Sample .par file

The par file contains one prior which is not individually replaced, such as expansion magnitude (under historical events) and will apply to all populations.

Then we use the following R-script and abc.R to do the final 1,000 acceptance and parameter estimation using local linear regression.

BayeSSC (Bayesian Serial SimCoal)
http://web.stanford.edu/group/hadlylab/ssc/