PINH

PINH: R programs for functional evaluation of the likelihood of paternal and maternal
recombination factions for pedigree data

1. INTRODUCTION

PINH is R programs for functional evaluation of the likelihood of paternal and maternal recombination factions for pedigree data. It contains all functions necessary for evaluating likelihood polynomial based on probability inheritance algorithm.

2. INSTALLATION

2.1 Installation of shared object

Some of the functions in PINH require C routines for rapid evaluation of the likelihood polynomial; therefore dynamic loading of a shared object is always required. Download appropriate shared object (pinh.so for 32-bit Unix system and pinh.dll for 32-bit Windows ) on your working directory, and install by

     > dyn.load("pinh.so")        # for 32-bit Unix system
     > dyn.load("pinh.dll")       # for 32-bit Windows

If dynamic loading is not successful or another operating system is used, download original source code (pinh.c) and built by yourself as follows:

     # R CMD SHLIB [options] [-o dllname] pinh.c

2.2 Installation of PINH programs

All of the R functions can be installed by using:

     > source("http://sugaya.datascience.jp/PIA/PINH/pinh.R")

3. INPUT

The main function of the program

pinh

accepts a list of three data frames, including a pedigree structure, disease status, and marker data, as well as a list of marker allele frequencies and disease allele frequency with penetrance as an attribute. The auxiliary functions,

read.prefile

read.datafile

, convert linkage pedigree input file or linkage parameter file to 2 lists, one of which is a list of the 3 data frames, and the other is a list of allele frequencies. One of the easiest ways to prepare the inputs is to prepare linkage pedigree input file and linkage parameter file, and then, converting them to 2 list objects as follows:

     > data = read.prefile("prefile.dat")
     > allele.prob = read.datafile("datafile.dat")

Otherwise, it is necessary to prepare 2 lists manually.

data: a list of the following 3 data frames.
- pedigree:
  a data frame with 5 variables; Pedigree.ID, ID, Father.ID, Mother.ID, and Gender. Father.ID and Mother.ID takes 0 if the ID is founder. Male is indicated by "M" and female by "F" in Gender variable.
- disease:
  a data frame with 3 variables; Pedigree.ID, ID, and Status. The Status takes NA if the affected status is unknown.
- marker:
  a data frame with 5 variables; Pedigree.ID, ID, Name, Allele1, and Allele2. The Name is the name of marker locus to be analyzed. Allele1 and Allele2 represent alleles at the marker locus.
allele.prob: a list of marker allele frequencies and disease allele frequency with penetrance as an attribute.
- marker:
  a vector of marker allele frequencies. This vector has name as same name as indicated in Name variable of marker data frame (see above.)
- disease:
  a vector of disease allele frequencies. It has penetrance vector as as an attribute.

4. EXECUTION

After creating 2 list objects, the main function,

pinh

, can be executed as follows:

     > pinh(data, allele.prob)

5. DOWNLOAD

PINH programs (pinh.R)
Shared object for 32-bit Unix system (pinh.so)
Shared object for 32-bit Windows (pinh.dll)
Original source code (pinh.c)

6. PUBLICATION

Yuki Sugaya and Ritei Shibata. Probability inheritance algorithm and its implimentation. Journal of Statistical Computation and Simulation, in press.