I am looking for a sample algorithm or a script for Mendelian inheritance filtering. I have a table of variants and each row represents a unique variation that is shared or not between different members of a large family. I imagine that there should be a meta file I supply with the Mendelian structure information and my script should read that and filter my data based on that meta data. The result should be these variants fit an autosomal recessive inheritance patterns etc. I can provide the expected inheritance pattern in the meta data file as well but the ideal script should be able figure that out as well. Does anyone know anything similar to this before I start melting my brain on this hopefully not reinventing the wheel? Python preferred but a general algorithm is more than OK.
Thanks
People around my neck of the woods tend to use PLINK
PLINK is still kept reasonably up to date and has good support for data formats/subsetting of data, summary statistics and population stratification detection - I think these two cover what you're asking. Certainly seems to offer much of what PedCheck/PedStats do as well.
There's also an additional Java based program to support some of the functionality and R plugins can be used to extend the programmatic functionality through Rserve
PedCheck is the standard Mendelian testing program that I've seen used in many studies, and distributed with other genetic analysis programs as a data quality tool. The paper describes the basic tests and examples. It is available after you register.
An alternative and more recent program is PedStats.
It changes the 1 based coordinates to 2 based coordinates after you read the file in it, so when going back to check predictions etc, you may want to reconvert. The Custom-query section is what you are looking for. There is a video of the author Jamie K. Teer showing the nuts and bolts from Mullikin lab: http://bit.ly/GT0BKx
