Familias is a free software for probability calculations when inferring paternity and identification based on DNA data. The latest version of the software as well as other useful files can be found in the Download section.
The Familias program may be used to compute probabilities and likelihoods in cases where DNA profiles of some people are known, but their family relationship is in doubt. The program extends a previous program known as pater, in that multiple alternative pedigrees can be handled. Given several alternative family trees (or pedigrees) for a group of people, given DNA measurements from some of these people, and given a data base of DNA observations in the relevant population, the program may compute which pedigree is most likely, and how much more likely it is than others. Obviously, there are several other programs performing similar tasks. As far we know a distinguishing feature of Familias is its ability to handle complex cases where potential mutations are accounted for, together with its ability to handle multiple pedigrees simultaneously. Familias does not co computations for linked markers; for such computations, you may consider FamLink.
The Familias software (and previously the pater program) has been developed by Petter Mostad and Thore Egeland, then at the Norwegian Computing Center, in cooperation with Bjørnar Olaisen, Margurethe Stenersen, and Bente Mevåg at the Institute of Forensic Medicine (RMI) in Oslo. The Norwegian Computing Center is a non-profit institute for applied research. The cooperation with RMI started in 1994, and the first version of the release was made in 1995. Several updates have been made since then, the latest in May 2010. The continued development is now undertaken by Daniel Kling at the Norwegian Institute of Public Health, Department of Family Genetics.
The simplest cases in which the Familias software can be used are paternity cases, i.e., cases where one wants to determine whether a man is the father of a certain child. In general, one would have DNA data for the child, the mother, and the putative father. One would also need data bases with the frequencies for the observed alleles in the relevant population. Given this, the program will calculate the odds for the putative father being the real father.In such paternity cases, one may say there are two alternative hypotheses, i.e., two alternative pedigrees: One where the man is the father, and one where he isn't. The strength of Familias is that it may calculate odds for any pair of alternative pedigrees. Thus, it may be used to calculate whether an uncle is indeed an uncle, whether two girls are indeed sisters, and so on. In fact, multiple different hypotheses may be handled.
In some systems used for human DNA typing, mutations are fairly frequent. If the possibility of mutations is not considered, one may mistakenly exclude a pedigree that turn out to be the correct one after considering all the data. The Familias software has the ability to weigh the possibility of mutations against other evidence.
A number of other features, such as the possibility to adjust for kinship, are also included. See the Documentation for further details.
The software may be downloaded here. A short tutorial, and a manual, are available. The program runs under Windows, and is intended for practical use in DNA testing labs. It is the result of many many years of cumulative development.
For research purposes, the core functions of Familias have now been made availiable in an R package. Please see the OpenFamilias pages.
If you have questions, please consult the Help section or contact Daniel Kling.
Egeland, Thore; Mostad, Petter; Mevåg, Bente; Stenersen, Margurethe: "Beyond traditional paternity and identification cases. Selecting the most probable pedigree." Forensic Science International Vol 110, Nr. 1, 2000.
Some other references
Egeland, Thore; Mostad, Petter: "Statistical Genetics and Genetical Statistics: a Forensic Perspective". Scand. Journal of Statistics, Vol 29, 2002.