All functions |
|
---|---|
Returns adjusted sebetahat's based on t likelihood so that we can use a normal likelihood. |
|
Use control genes to estimate hidden confounders and variance inflation parameter, then run ASH. |
|
Use control genes to estimate hidden confounders and variance inflation parameter, then run ASH. |
|
Wrapper for |
|
The Evidence lower bound. |
|
Fixed point iteration for BACKWASH. |
|
Objective function for BACKWASH. |
|
Update for the "g" hyperparameter. |
|
Update for the prior mixing proportions. |
|
Update for the variational density of beta |
|
Update for the variational density of v. |
|
Update for the variance scaling parameter. |
|
BACKWASH: Bayesian Adjustment for Confounding Knitted With Adaptive SHrinkage. |
|
Second step of the backwash procedure. |
|
Fast Gibbs sampler for Bayesian factor analysis. |
|
Bayesian factor analysis used in Gerard and Stephens (2021). |
|
Simple Bayesian low rank matrix decomposition. |
|
Fast Gibbs sampler for Bayesian factor analysis. |
|
R implementation of |
|
Wrapper for the bfa package. |
|
Simple Bayesian low rank matrix decomposition. |
|
Objective function for updating variance inflation parameter during EM for mixtures of normals prior. |
|
Gibbs sampler for Bayesian SVD. |
|
Empirical estimate of lfsr based on posterior samples. |
|
Same as |
|
Calculate moments of pointmass rv's. |
|
Calculate quantiles of pointmass rv's. |
|
Iterative procedure for confounder correction with a procedure that returns lfdrs. |
|
Use the distribution of control genes' p-values to adjust all of the p-values. |
|
Quantile normalize stats to their theoretical distributions. |
|
RUV4's second step. |
|
RUV4's second step. |
|
Density of a mixture of normals |
|
Wrapper for dt with a non-zero mean and non-1 scale parameter. |
|
EM algorithm for factor analysis with missing block matrix. |
|
Fixed point iteration for em algorithm with missing block. |
|
Faster version of |
|
The objective function for em with a missing block. |
|
A faster version of em_miss_obj. |
|
Wrapper for cate's fa.em function. |
|
Test output form fa_func. |
|
Tests whether a user-specified factor analysis function
is appropriate for use in |
|
Test whether a user-specified function is compatible with |
|
A fixed point iteration for CAW. |
|
A wrapper for |
|
Old version of Bayesian factor analysis. |
|
Default way to set grid of variances. |
|
My version of hard imputation that begins at the ruv estimates |
|
Hierarchical prior density function as described in Gerard and Stephens (2021) |
|
Constructs an overall matrix, then applies a given imputation function. |
|
Reproduce RUV2, RUV3, and RUV4 with RUVimpute. |
|
Function for initializing mixing proportions. |
|
Wrapper for impute.knn |
|
Wrapper for missForest package. |
|
MOUTHWASH: Maximize Over Unobservables To Help With Adaptive SHrinkage. |
|
Coordinate ascent for optimizing t likelihood with uniform mixtures. |
|
The second step of MOUTHWASH. |
|
Gradient wrt z of |
|
A fixed point iteration for updating the mixing proportions and the confounders associated with the covariates of interest when using a mixture of normals prior. |
|
Wrapper for |
|
Penalized MOUTHWASH likelihood when using a mixture of normals. |
|
Wrapper for |
|
A basic normal prior density function. |
|
PCA when first |
|
Basic PCA. |
|
PCA when first |
|
Fix point for mle in pca_ruv2. |
|
The objective function for mle in pca_ruv2. |
|
Plotting method for |
|
Plotting method for |
|
Plotting method for |
|
Wrapper for |
|
Wrapper for pt with a non-zero mena and non-1 scale parameter. |
|
More stable way to calculate differences in T cdf's. |
|
More stable way to calculate differences in T cdfs when input is a matrix. |
|
Log version of |
|
Objective function for quasi-mle approach. |
|
Basic no optimized code objective function. |
|
Quasi-mle when first |
|
Quasi-mle when first |
|
An identified QR decomposition. |
|
Draw from a mixture of normals. |
|
Random draw from a mixture of normals. |
|
QR rotation to independent models. |
|
Very inefficient copy of Rubin and Thayer iteration mostly meant for debugging. |
|
Removing Unwanted Variation 3. |
|
Bayesian version of Removing Unwanted Variation. |
|
Same as ruvimpute but only does ruvem and comes up with estimates of standard errors. |
|
General imputation framework. |
|
A simulated RNA-seq dataset. |
|
Wrapper for a Kolmogorov-Smirnov test using the p-values from the control genes. |
|
Wrapper for a Kolmogorov-Smirnov test using the statistics from the control genes. |
|
A wrapper for using the softImpute function from the softImpute package. |
|
EM algorithm to find regression coefficients using t-likelihood when variances are known up to scale. |
|
Fixed point iteration for EM algorithm for regression with t-errors where the variance is known up to scale. |
|
The likelihood for regression with t-errors where the variance is known up to scale. |
|
Wrapper for |
|
Gradient wrt z2 of intermediate function in EM. |
|
Intermediate objective function in EM. |
|
Fixed point iteration when using t errors and a mixture of uniforms prior. |
|
Wrapper for |
|
Log-likelihood when using t errors and mixture of uniforms prior. |
|
Wrapper for |
|
Utility function for updating F in |
|
Wrapper for |
|
Fast update for sig_diag and alpha. |
|
Basic no optimized code update for sig_diag and alpha. |
|
|
|
Calibrated RUV2. |
|
Old version of Calibrated RUV2 that doesn't really work. |
|
Calibrated RUV4 where the control genes are used to estimate hidden confounders and a variance inflation parameter. |
|
Variance inflated RUV-inverse. |