Mendel

Online Docs :: Download

The Mendel software program performs likelihood-based statistical analysis to solve a variety of genetic problems. Implementations are included for all common, and several novel, statistical genetic tests. Data sets can consist of qualitative or quantitative traits, pedigree or population samples, limited loci or dense SNPs. Where appropriate, the analysis will use either the Elston-Stewart or Lander-Green-Kruglyak algorithms, whichever is more efficient for each individual pedigree.

Mendel includes options to:

  1. order and map a set of markers along a chromosome;
  2. use parametric linkage methods to localize a gene on a fixed marker map;
  3. use non-parametric linkage (NPL) methods to localize a gene on a fixed marker map;
  4. estimate pedigree haplotypes;
  5. identify potential genotyping errors;
  6. test for allelic association by the TDT or the gamete-competition model;
  7. determine kinship coefficients conditional on marker data;
  8. perform genetic risk calculations;
  9. test for paternity or other pedigree relationships;
  10. estimate allele frequencies using either pedigrees or a random sample;1
  11. estimate trait genotype penetrances;
  12. estimate ethnic composition using pedigree data;
  13. identify deviations from Hardy-Weinberg and linkage equilibrium;
  14. simulate genetic data by gene dropping;
  15. localize QTLs, including X-linked QTLs, using variance component analysis;
  16. perform association analysis on quantitative traits;
  17. test for association given linkage; and
  18. trim pedigrees to a core group and all necessary connecting relatives.
  19. dense SNP genotype imputation, including haplotypes.
  20. GWAS: dense SNP association testing, including rare variants.
  21. maternal-fetal genotype incompatibility test.
  22. inbred strain QTL analysis.
  23. simulate trait values incorporating major loci and various models.
  24. pedigree-based GWAS (SNP association testing).

Contact and Citation Information

Contact Information:

Any comments, suggestions, and questions regarding Mendel are welcome at mendel@genetics.ucla.edu.

Citation:

Lange K, Papp JC, Sinsheimer JS, Sripracha R, Zhou H, Sobel EM (2013) Mendel: The Swiss army knife of genetic analysis programs. Bioinformatics 29:1568-1570.

New Features

New Features in Mendel Version 13

This new version has significantly expanded several existing options and add one entirely new analysis option. The new Ped-GWAS analysis option provides extremely fast GWAS analysis on extended pedigrees, nuclear families, unrelated individuals, or any combination thereof. It can handle univariate or multivariate quantitative traits with missing data, and allows for covariate adjustment, including correction for population stratification. Pedigree kinships can either be explicitly provided or automatically estimated from dense markers. SNPs can be analyzed under codominant (additive), dominant, or recessive models.

The Ethnic Admixture analysis option has been expanded to rank markers by their informativeness to distinguish subpopulations. This option can now also perform very efficient principal component analysis (PCA). These two new features can be used with either text-based or binary data files.

The Kinship analysis option has also been expanded to binary files. It can now quickly estimate SNP-based global and local kinship coefficients on dense genomewide data. It can also use the SNP-based global estimates to cluster individuals and thus construct pedigree groupings without pre-existing relationship information.

The input data formats are now more flexible. For example, PLINK binary data sets (.fam, .bin, and .bed files) are now accepted (see the Mendel Documentation for details). The SNP definition file can now include allele names. Analogous to the SNP subset file, there is now a sample subset file to easily change which subset of individuals should be included in an analysis. Also, pedigree files without the twin field or without any pedigree fields can be read directly. These options are set using the new keyword INPUT_ FORMAT.

Many smaller improvements were also implemented. For example, missing SNP genotypes are now replaced with random ``average'' genotypes when missing data is not allowed in an analysis. Each selection of a random genotype is based on the allele frequencies within the data set. The GWAS output has been expanded and an issue with the Q-Q plot data is fixed. Also in the GWAS analysis option, one can now set specific values for the LASSO tuning constants using the command TUNING_ CONSTANT, if one wishes to override Mendel's automatic adjustments.

In version 13.2 the analysis to rank markers by their informativeness to distinguish subpopulations, has been extended to allow non-uniform population contributions. There are also a few minor bug fixes and other improvements.

New Features in Mendel Version 12

In this update we have added one new option, Trait Simulation, and improved several of the existing options. The new option uses either existing genotype data or the output of the genotype simulation option to determine trait values. The traits can be either univariate (simulated by generalized linear models) or multivariate (simulated by variance component models). Mean effects due to the genotypes at a major locus can be included in the model. The list of supported distributional families for the mean effects model is substantial (see Tables 14.1 and 14.2 in the Documentation).

Genotype simulation now allows various output styles. In addition, one has considerable control over the degree of missing data for both trait and genotype simulation. This flexibility is accomplished with the introduction of four new keywords: GENE_ DROP_ OUTPUT, KEEP_ FOUNDER_ GENOTYPES, MISSING_ AT_ RANDOM, and MISSING_ DATA_ PATTERN.

Since a grand mean (intercept) is required for each trait in all of Mendel's regression analyses, it is now added automatically. Thus it is no longer necessary to add commands such as PREDICTOR = GRAND :: trait to your model. It does no harm if you continue to include such commands explicitly.

An important change has been made to the meaning of the weight that can be assigned to each SNP in the SNP definition file. These predictor weights are used when building the best regression model that includes the user-specified number of predictors. A weight can now be any positive real number. Also, the greater the weight, the more likely the predictor will be included in the model. If the keyword UNIFORM_ WEIGHTS has the value True, which is the default, then all predictors with unassigned weights receive weight 1.0. Otherwise, SNPs with unassigned weights receive weight 1/sqrt{4q(1-q)}, where q is the minor allele frequency at the SNP. (If you previously included weights in your SNP definition files, we suggest you now use the inverse of the old weights.)

Other minor changes include the maximum number of loci that can be combined into a superlocus has been increased from four to five. The keyword SEED can now be set to the special value TIME, which sets the seed based on the current time. The value set is reported in the standard output file.

New Features in Mendel Version 11

This revision incorporates two new options and several improvements to existing options. The new options are Maternal-Fetal Genotype Incompatibility Testing and Inbred Strains Analysis. The first of the new options facilitates modeling of interactions between maternal and child genotypes. The classic example is Rh maternal-fetal incompatibility. The second of the new options represents a new approach to QTL mapping with inbred strains of mice and other animals. The option implements a mixed effects model that correctly captures polygenic background, handles multivariate traits, and copes with pedigrees of arbitrary complexity. The QTL effect is modeled at the mean level as a vector of regression coefficients on the strain origin pair (maternal-paternal) imputed for each animal at the current QTL location along the genome.

To better deal with rare SNPs, the SNP Association option has been significantly improved. In an expanded version of the SNP Definition file, for each SNP one can now assign a weight and membership in a group of SNPs, e.g., a gene or molecular pathway. Default weights can be based on allele frequencies. The lasso regression can now use group penalties, which make it easier to understand association at the gene level rather than the SNP level. The new keyword LASSO_ PROPORTION specifies the relative importance the model puts on individual predictors versus predictor groups. One may also specify individual predictors, or groups of predictors, to always retain in the LASSO model by using the keywords RETAINED_ PREDICTOR and RETAINED_ GROUP. Finally, in interaction testing, it is now possible to test for all pairwise interactions in a large base set of predictors.

Among the modification of existing features, we have introduced a better facility for imposing linear constraints on parameters. See the discussion of the new keyword PARAMETER_ EQUATION in the documentation.

The SNP Imputation option now performs both pedigree based and linkage disequilibrium based genotype and haplotype inference. The combination of these two approaches leads to more accurate imputation results. Better imputation in turn leads to better handling of missing data and presumably more sensitive SNP association tests.

Mendel 2013 Software License Agreement

Please read and specify that you agree or disagree at the bottom of this document.

Name of software being licensed: MENDEL

THIS IS A LEGAL AGREEMENT BETWEEN YOU (EITHER AS AN INDIVIDUAL OR ENTITY) AND KENNETH L. LANGE. BY INSTALLING THIS SOFTWARE, YOU ARE CONSENTING TO BE BOUND BY AND ARE BECOMING A PARTY TO THIS AGREEMENT. IF YOU DO NOT AGREE TO ALL OF THE TERMS OF THIS AGREEMENT, DO NOT INSTALL MENDEL.

THIS LICENSE COVERS ONLY THE YEAR 2013 AND THE FIRST FOUR MONTHS OF THE FOLLOWING YEAR. AT THE EXPIRATION OF THIS LICENSE, YOU MUST DESTROY ANY COPIES OF MENDEL OBTAINED UNDER THIS LICENSE. IF YOU WISH TO CONTINUE USING MENDEL, THEN YOU MUST DOWNLOAD A NEW COPY AND AGREE TO A NEW LICENSE.

TERMS OF THE LICENSE:

FEES: You may download Mendel as often as necessary, free of charge.

RESTRICTIONS: You may NOT

(a) copy, redistribute, post, or otherwise enable or permit other individuals to access or use MENDEL except under the terms listed herein;

(b) permit concurrent use of the Software except under the terms listed herein;

(c) modify, translate, reverse engineer, decompile, disassemble, or create derivative works based on MENDEL;

(d) rent, lease, grant a security interest in, or otherwise transfer rights to Mendel; or

(e) remove or put any proprietary notices or labels on MENDEL.

TITLE: Title, ownership rights, and intellectual property rights in MENDEL shall remain with Kenneth Lange. MENDEL is protected by copyright and other intellectual property laws and by international treaties.

DISCLAIMER OF WARRANTY: MENDEL IS PROVIDED AS IS WITHOUT WARRANTY OF ANY KIND. KENNETH LANGE MAKES NO REPRESENTATIONS AND EXTENDS NO WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT. THE ENTIRE RISK ARISING OUT OF THE USE OR PERFORMANCE OF MENDEL AND ITS DOCUMENTATION REMAINS WITH LICENSEE. IN NO EVENT SHALL KENNETH LANGE BE LIABLE FOR ANY CONSEQUENTIAL, INCIDENTAL, DIRECT, INDIRECT, SPECIAL, PUNITIVE, OR OTHER DAMAGES WHATSOEVER (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS INFORMATION OR OTHER PECUNIARY LOSS) ARISING OUT OF THIS AGREEMENT OR THE USE OF OR INABILITY TO USE MENDEL.

INSTALLATION: LICENSEE warrants and represents that a competent System Administrator will install MENDEL, and that the System Administrator will carefully review all documentation, manual pages, and release notes included with MENDEL prior to installation.

USE: LICENSEE agrees to provide KENNETH LANGE with a report of any errors encountered in LICENSEE'S use of MENDEL.

RISK: The entire risk as to the use and performance of MENDEL is assumed by LICENSEE. LICENSEE shall defend, indemnify and hold harmless KENNETH LANGE AND HIS EMPLOYER, THE UNIVERSITY OF CALIFORNIA, for and against any and all claims, demands, damages, losses, and expenses of any kind (including but not limited to attorneys' fees), relating to or arising from any use or disposition by LICENSEE or its transferees of MENDEL.

MAINTENANCE: Neither KENNETH LANGE nor the UNIVERSITY OF CALIFORNIA is obligated to provide maintenance, technical support, interpretation of results, or updates for MENDEL.

TERMINATION: This license will terminate automatically if you fail to comply with the limitations described herein. On termination, you must destroy all copies of the Software and Documentation.

ACKNOWLEDGEMENT: LICENSEE agrees to reference MENDEL in every publication that employs it. See the top of every standard output file generated by MENDEL for the preferred citation to use.