Sequential Importance Sampling To Estimate Relatedness Structures (SISTERS)

This code implements my sequential importance sampling (SIS) methodology for making inference about the pedigree, or relatedness, between a set of genotyped individuals.

Compilation

Type make. At the moment, only a Makefile for GNU/Linux type systems is supplied. The code links against the GSL.

Input

To run the program, type ./sisters [switches] [name-of-input-file]

The input file should contain the genotypes of the sampled individuals, in the same format as used by the popular PHASE program. See the included example input file.

The following can also be specified as command line switches

-e Assume the Enjalbert-David model for ancestors

-g Assume gonochoristic (dioecious) mating system instead of hermaphrodites

-h Assume ancestors in exact Hardy--Weinberg

-nNUMBER Set number of ancestors/candidate parents

-pNUMBER Set number of particles to use in SIS

-r Use residual resampling

-v Generate verbose output

-xNUMBER Set number of replicate runs

The approach is more fully described in a recent presentation.

Output

The program prints progress information and convergence diagnostics to the standard output as it runs.

To pairwise probabilities of full/half sibship are output in a file called pairprobabilities, and the individual probabilities of being products of selfing are output in a file called selfstatus.

Download

Click here to download an alpha version (version 0.11) of the code.