Famtypes

Famtypes
A simple utility for parsing PED files	Shaun Purcell, Massachusetts General Hospital

Introduction

What does Famtypes do?

Checks for data entry errors:

Checks that the pedigree file has the expected number of columns.
Checks that everyone listed as a parent has an entry. Creates missing entries by assuming the missing person is a founder, unavailable, and has unknown affection status, phenotypes and genotypes.
If only one parent is listed, creates an imaginary parent. This ensures everyone, except founders, has two parents. Famtypes sets these dummy entries as founders, with unknown affection status, phenotypes and genotypes. It lists all dummy entries created in the log file.
Lists half-siblings in the log file.
Unknown genders are filled in from offspring data, where possible (if all offspring list the parent as a single gender) and changed to male if not (no offspring or conflicting description of parent gender). All changes are listed in the log file.
If a family is disjoint (has people who are not related in any way) Famtypes breaks that family into groups of people who are related and list the changes in the log file. (If there are only individuals with no parents or offspring, the user can decide to assume these are siblings, and Famtypes will create dummy parents for them.)

Filters and reformats data:

The user can choose to break multigenerational families into multiple nuclear families.
The user can choose to filter the data so that only families of a certain type are listed in the log file statistics. They can be filtered by:
1. All families. (no filter)
2. All families with both parents and at least 1 offspring available.
3. All families with both parents and at least 1 affected offspring available.
4. The nuclear family with the most affected offspring in a multigenerational family, after ignoring unavailable people. (This means only families with two available parents and at least one available and affected offspring will be considered.)
Note: If affection status or availability columns are not provided, all individuals will be treated as available and affected by the filters.
Returns a cleaned up pedigree file and a text file of statistics and changes between the inital pedigree file and the cleaned up version.

Download

Famtypes is in beta, which means we're still working on it, adding features and fixing problems as we find them. Check the warnings section for known problems, and use at your own risk...

People within the MGH network can also run Famtypes on Pnguapp, by following this internal link. Others should download a copy onto their own machine using the links below.

Famtypes runs in DOS, Linux and OS X. The source code is also available for people who want to compile it themselves. Just right-click the appropriate link and choose 'Save-As...':

Files:

Version 1.4: Fixed availability column requirement, fixed bugs in command-line version.
Version 1.3: Added available individual and family counts. Forced use of availability col temporarily, added version and parameter notes to log file.
Version 1.2: Fixed filter 4, added check for duplicate indIDs within a family.
Version 1.1: Fixed avail, added ability to run from command line, compiles on machines tested.
Version 1.0: Known bug - required availability column, only compiled on some machines

MS-DOS/Windows executable file: dosft1_4.exe
Linux binary file: linft1_4.tar.gz
Mac(intel) binary file: mac_intel_famtypes1.4.tar.gz
C/C++ source: sourceft1_4.tar.gz

Installation:

MS-DOS/Windows: The .exe file can be saved to a folder or directly to the Desktop. To run interactively, just double-click on the icon.
Linux and OS X: download the binary to a a folder or directory. To run interactively, type ./famtypes while in that directory.
From source: Can compile on any platform that supports C/C++ see notes for similar program.

Please e-mail me if you download a copy, especially if you encounter any problems using Famtypes or have suggestions.

Pedigree File Format

Each row in a pedigree file has information for a single person. The first five columns, in order, need to have a family ID number (famID), an individual ID number within the family (indID), the indID of the person's father (patID), indID of mother (matID), and gender.

family ID number (famID)
individual ID number (indID, must be unique within a family)
father's indID (If founder, set fatID to 0)
mother's indID (If founder, set matID to 0)
gender (male - 1, female -2)

Below is a sample pedigree file for two nuclear families. Family 1 consists of a father (indID 1), mother (indID 2) and male child (indID 3). Family 2 consists of a father (indID 1), mother (indID 2), female child (indID 3) and male child (indID 4).

Famtypes will also ask about several optional columns. Any columns that are used should be in the following order:

affection status (unknown - 0, unaffected - 1, affected - 2)
availability (unavailable - 0, available - 1)
phenotype(s) (can by 0-2 or continuous)
genotype(s) (Each genotype takes 2 columns. This data is usually encoded as letters (A,T,G,C, x=unknown) or numbers (1-4, 0=unknown), but any encoding with two columns is fine.)

Any individual created by Famtypes will have 0 in both affection status and availabilty columns if those columns exist in the pedigree file. If these columns are not being used, individuals created by Famtypes will be treated by filters as available and affected, just like the other individuals in the pedigree file.

Here, the data from the first example have been extended to include affection and availability status, 2 phenotypes and 2 genotypes.

1 1 0 0 1   0 0   0 0.0   0 0   0 0
1 2 0 0 2   1 1   1 3.0   3 4   3 4
1 3 1 2 1   2 1   0 1.5   3 3   3 3

Examples

Example 1: Missing vs Unknown Parent.

In this example a mother is missing from the pedigree file. In exMissPar.ped, the mother's ID is known since the offspring list the matID as 2. In exUnknPar.ped, the offspring have matID of 0, for unknown mother.

For the missing parent, Famtypes adds in a person to match the maternal ID referenced. You can download the log of changes, and the resulting pedigree file, which is shown below:

1 1 0 0 1   0 1
1 2 0 0 2   0 0
1 3 1 2 2   0 1
1 4 1 2 1   0 1

Compare this to the result of an unknown parent. Here, both offspring list a matID of 0, for unknown mother. In this case we can't assume they both have the same mother, since they might be half-siblings. log of changes, and the resulting pedigree file, which is shown below:

MG1_2   1   0   0   1   0 1
MG1_2 DUMM1 0   0   2   0 0
MG1_2   3   1 DUMM1 2   0 1
MG2_2   1   0   0   1   0 1
MG2_2 DUMM2 0   0   2   0 0
MG2_2   4   1 DUMM2 1   0 1

Example 2: Disjoint Families

In this example, two unrelated trios are listed as part of the same family.

Famtypes splits them into two different families by adding a group number (GRP) to the front of the famID.

GRP1_1 1 0 0 1
GRP1_1 2 0 0 2
GRP1_1 3 1 2 2
GRP2_1 4 0 0 1
GRP2_1 5 0 0 2
GRP2_1 6 4 5 1

Usage

First, you need access to a copy of Famtypes. The download section gives several sets of instructions for doing this.

There are two ways to run Famtypes, either interactively or from the command line.

Run Interactively:

Open a terminal
Change to the Famtypes directory.
(Type something like cd user/famtypesFolder/famtypes)
Make sure the pedigree file (filename.ped) you want to run is in this directory.
(Type ls and make sure your file is listed)
run Famtypes
(Type ./famtypes)
Famtypes will ask the following questions about the file and how to process it:
- Name of file to analyze - filename.ped
- Is there an affection status column - y or n
- Is there an availability column - y or n
- Should Famtypes assume completely disjoint families are really siblings - y or n
- Number of phenotypes and covariates beyond column 5 - 0, 3, 15, etc
- Number of genotypes (1 genotype takes 2 columns) - 0, 1, 23, etc
- Name of new .ped output file - newfile.ped
- Name of log file/secondary output file - logfile.txt
- Type of filter (see intro for options) - 1, 2, 3, or 4
once it finishes, there should be two new files with the names Famtypes asked for. The first is a cleaned up version of the pedigree file and the second is a text file with statistics and a list of changes between the original and new pedigree files.

Run from the Command Line:

Follow the interactive instructions, but when running Famtypes, give all 9 answers to the interactive questions above as command line arguments. For example, in test.ped (shown below), thre is an affection status and availability status column, but no phenotypes or genotypes.

1 1 0 0 1  1 1
1 2 0 0 2  1 1
1 3 1 2 1  1 1

To run this from the command line without treating disjoint families as siblings and without filtering, type the following.

./famtypes test.ped y y n 0 0 testOut.ped testLog.txt 1

Warnings

Known problems:

If a pedigree file lists someone as their own parent, grandparent, etc, Famtypes will freeze in an infinite loop.

Future Developments

Citation & Contact

Famtypes was developed by

Shaun Purcell (Institute of Psychiatry, London, UK & Whitehead Institute, MIT, Cambridge, MA, USA)

Created by Shaun Purcell : Aug 2005.
Updated by Lori Thomas : Feb 2006.