Charles Kooperberg - Software

Genetic association studies

Code for fitting an ultra-sparse variational Bayes spike regression model, with applications to Genome-wide association studies (vBsr) is here. WARNING: Code is in an early release form (version 0.0.1), and details of the model are forthcoming. Email questions to Ben Logsdon (blogsdon@u.washington.edu).

Code for approximate power calculations for identification of gene x gene and gene x environment interactions in genomewide association studies using a two-stage analysis. Download the package powerGWASinteraction from CRAN.

This method is described in C. Kooperberg and M. LeBlanc (2008). Increasing the power of identifying gene x gene interactions in genome-wide association studies Genetic Epidemiology, 32, 255-263.

This method is described in M. LeBlanc and C. Kooperberg (2009). Adaptively Weighted Association Statistics Genetic Epidemiology, 33, 442-452.

This method is described in J. Y. Dai, M. LeBlanc, N. L. Smith, B. Psaty and C. Kooperberg (2009). SHARe: an algorithm to select the most informative set of SNPs for candidate genetic association. Biostatistics, 10, 680-693.

This method is described in J. Y. Dai, M. LeBlanc, and C. Kooperberg (2009). Semiparametric estimation exploiting covariate independence in two-phase randomized trials. Biometrics, 65, 178-187.

Logic Regression

Logic Regression.
Logic Regression: R libraries to fit logic regression models.

The Logic regression program is described in Ingo Ruczinski, Charles Kooperberg, and Michael L. LeBlanc (2003). Logic Regression. Journal of Computational and Graphical Statistics, 12, 475-511.

Extensions are discussed in Charles Kooperberg and Ingo Ruczinski. Identifying interacting SNPs using Monte Carlo Logic Regression. Genetic Epidemiology, 28 , 157-170.

Microarrays

Improved background correction for spotted DNA microarrays: R code. It is also implemented in Gordon Smyth's Limma package on bioconductor as the function kooperberg.

This method is described in C. Kooperberg, T. G. Fazzio, J. J. Delrow, and T. Tskuiyama (2002) Improved background correction for spotted DNA microarrays Journal of Computational Biology, 9, 55-66.

Directed indices for exploring gene expression data: code is here. (Code written by Mike Leblanc mleblanc@fhcrc.org.)

This method is described in M. LeBlanc, C. Kooperberg, T. Grogan, and T, Miller (2003) Directed indices for exploring gene expression data Bioinformatics, 19, 686-693.

Extreme regression code is here. (Code written by Mike Leblanc mleblanc@fhcrc.org.)

This method is described in M. LeBlanc, J. Moon, and C. Kooperberg. (2006) Extreme regression. Biostatistics, 7, 71-84.

Polynomial Splines

Libraries

Software for hare, heft, lspec, logspline, polyclass and polymars, as well as the logspline version discussed in Stone et al (1997) are now available as an R-package. You can download the package polspline from CRAN. (Version 1.0.11, 5/24/2005.)

Software for logspline as well as the logspline version discussed in Stone et al (1997) are now available as an R-package. You can download the package logspline from CRAN. (Version 2,9,2, 4/18/2005.) The latest version of logspline is always part of the latest version of polspline .

A book about splines

Individual programs

POLYCLASS.
Polyclass: S functions to estimate conditional class probabilities based on a discrete response variable and a number of predictors. tar.gz archive.
I suggest downloading polspline (see above), it will give you all programs at once.

POLYMARS.
Polymars: A set of S functions for classification and (polychotomous) regression using adaptively selected polynomial splines for the model building. (Written by Martin O'Connor under my direction.)
I suggest downloading polspline (see above), it will give you all programs at once.

The POLYCLASS and POLYMARS programs are described in Polychotomous Regression. Charles Kooperberg, Smarajit Bose and Charles J. Stone. Journal of the American Statistical Association , 92, , 117-127 (1997).

HEFT.
Hazard Estimation with Flexible Tails: S functions to estimate an unknown hazard function based on possibly censored data and obtain corresponding densities probabilities, quantiles and random samples.
I suggest downloading polspline (see above), it will give you all programs at once.

The HARE and HEFT programs are described in Hazard Regression. Charles Kooperberg, Charles J. Stone and Young K. Truong. Journal of the American Statistical Association , 90, , 78-94 (1995).

OLDLOGSPLINE.
S functions to, fully automaticly, estimate an unknown density based on possibly grouped or censored data and obtain corresponding probabilities, quantiles and random samples. This function used to be known as ``logspline'', it uses stepwise knot deletion from a larger number of initial knots.
I suggest downloading polspline (see above), it will give you all programs at once.
The OLDLOGSPLINE program is described in Logspline Density Estimation for Censored Data. Charles Kooperberg and Charles J. Stone. Journal of Computational and Graphical Statistics, 1, 301-328, (1992).

LOGSPLINE.
S functions to, fully automaticly, estimate an unknown density based on possibly grouped or censored data and obtain corresponding probabilities, quantiles and random samples. This function used to be known as ``nlsd'', it uses stepwise addition and stepwise deletion of initial knots.
I suggest downloading polspline (see above), it will give you all programs at once.
The differences between the LOGSPLINE program and OLDLOGSPLINE are described in The use of polynomial splines and their tensor products in extended linear modeling (with discussion). Charles J. Stone, Mark H. Hansen, Charles Kooperberg, and Young K. Truong. The Annals of Statistics, 25, 1371-1470, (1997).

The LSPEC program is described in Logspline Estimation of a Possibly Mixed Spectral Distribution. Charles Kooperberg, Charles J. Stone and Young K. Truong. Journal of Time Series Analysis, 16, 359-388 (1995).