simulation study of initial variance influence on model fitting

Last updated: 2023-02-24

Checks: 7 0

Knit directory: gsmash/

This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Environment: empty

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

Seed: set.seed(20220606)

The command set.seed(20220606) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Session information: recorded

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Cache: none

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

File paths: relative

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Repository version: 098c8a2

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version 098c8a2. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    analysis/figure/

Untracked files:
    Untracked:  analysis/compare_vga_iterations_pbmc_full.Rmd
    Untracked:  code/poisson_STM/get_loadings_order.R
    Untracked:  output/pbmc3k_k1.rds
    Untracked:  output/pbmc3k_k1_S1.rds

Unstaged changes:
    Modified:   analysis/overdispersed_splitting.Rmd
    Modified:   code/ebpmf_simulation/simu_func.R
    Modified:   code/poisson_STM/plot_factors.R
    Modified:   code/poisson_STM/plot_factors_general.R
    Modified:   code/poisson_STM/structure_plot.R

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.

These are the previous versions of the repository in which changes were made to the R Markdown (analysis/simu_study_of_init_variance.Rmd) and HTML (docs/simu_study_of_init_variance.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File	Version	Author	Date	Message
Rmd	098c8a2	DongyueXie	2023-02-24	wflow_publish("analysis/simu_study_of_init_variance.Rmd")

Introduction

So in a previous study, I found that different initial variance can lead to different structures.

In this study, I ran the following simulation:

A. Use the fitted model that inits with small variance to generate data. Specifically, the 1e-6, 60th iter model.

B. Use the fitted model that inits with larger variance to generate data. Specifically, the 1e-2, 60th iter model.

For both A,B, I tried to fit ebpmf model with small and large init variances. Each experiment is repeated 5 times.

library(fastTopics)
library(Matrix)
library(gridExtra)
source('code/poisson_STM/plot_factors.R')
source('code/poisson_STM/plot_factors_general.R')
source('code/poisson_STM/structure_plot.R')
source('code/poisson_STM/get_loadings_order.R')
data("pbmc_facs")
cell_names = pbmc_facs$samples$subpop

Small true variance, small init variance

res = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc3k_simulation/simu_pbmc_fasttopics_small_true_var_small_init_var.rds')
K_hat = c()
for(i in 1:length(res$output)){
  K_hat = rbind(K_hat,c(res$output[[i]]$fitted_model$flash$n.factors,res$output[[i]]$fitted_model$ebpmf$fit_flash$n.factors))
}
colnames(K_hat) = c('flash','ebpmf')
K_hat

     flash ebpmf
[1,]     5     8
[2,]     5     8
[3,]     5     8
[4,]     5     8
[5,]     5     8

plot0=plot.factors.general(res$sim_data$Loading,
                           cell.types=cell_names,title='True Loadings',
                           nonnegative = T,kset = 3:8)
for(i in 1:length(res$output)){
  plot1 = plot.factors(res$output[[i]]$fitted_model$flash,
                       cell.types=cell_names,title='flash',
                       nonnegative = T)
  plot2 = plot.factors(res$output[[i]]$fitted_model$ebpmf$fit_flash,
                       cell.types=cell_names,title='ebpmf',
                       nonnegative = T,rm_kset = c(1,2))
  grid.arrange(plot1, plot0,plot2, ncol=3)
}

loadings_order = get_loadings_order(res$sim_data$Loading,res$sim_data$Factor,grouping = cell_names,n_samples = 5000)

Perplexity automatically changed to 53 because the original setting of 100 was too large for the number of samples (163)

plot0=structure_plot_general(res$sim_data$Loading,res$sim_data$Factor,grouping =cell_names,title = 'True',print_plot = F,loadings_order=loadings_order)

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

for(i in 1:length(res$output)){
  plot1 = structure_plot_general(res$output[[i]]$fitted_model$flash$L.pm,
                                 res$output[[i]]$fitted_model$flash$F.pm
                                 ,grouping=cell_names,
                                 title='flash',
                                 print_plot = F,
                                 loadings_order=loadings_order,
                                 remove_l0f0 = F)
  plot2 = structure_plot_general(res$output[[i]]$fitted_model$ebpmf$fit_flash$L.pm,
                                 res$output[[i]]$fitted_model$ebpmf$fit_flash$F.pm,
                                 grouping =cell_names,
                                 title='ebpmf',
                                 print_plot = F,
                                 loadings_order=loadings_order)
  grid.arrange(plot1, plot0,plot2, nrow=3)
}

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

fit = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc_fasttopics/init_tol_effect/nonnegLF_pe_inittol1e6_iter60.rds')
for(i in 1:length(res$output)){
  plot(fit$sigma2,res$output[[i]]$fitted_model$ebpmf$sigma2,col='grey80',pch=20,xlab='true',ylab='estimated',main='variance estimation')
  abline(a=0,b=1)
}

Small true variance, large init variance

res = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc3k_simulation/simu_pbmc_fasttopics_small_true_var_large_init_var.rds')
K_hat = c()
for(i in 1:4){
  K_hat = rbind(K_hat,c(res$output[[i]]$fitted_model$flash$n.factors,res$output[[i]]$fitted_model$ebpmf$fit_flash$n.factors))
}
colnames(K_hat) = c('flash','ebpmf')
K_hat

     flash ebpmf
[1,]     5     8
[2,]     5     8
[3,]     5     8
[4,]     5     8

loadings_order = get_loadings_order(res$sim_data$Loading,res$sim_data$Factor,grouping = cell_names,n_samples = 5000)

Perplexity automatically changed to 53 because the original setting of 100 was too large for the number of samples (163)

plot0=structure_plot_general(res$sim_data$Loading,res$sim_data$Factor,grouping =cell_names,title = 'True',print_plot = F,loadings_order = loadings_order)

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

for(i in 1:4){
  plot1 = structure_plot_general(res$output[[i]]$fitted_model$flash$L.pm,
                                 res$output[[i]]$fitted_model$flash$F.pm
                                 ,grouping=cell_names,
                                 title='flash',
                                 print_plot = F,
                                 loadings_order=loadings_order,
                                 remove_l0f0 = F)
  plot2 = structure_plot_general(res$output[[i]]$fitted_model$ebpmf$fit_flash$L.pm,
                                 res$output[[i]]$fitted_model$ebpmf$fit_flash$F.pm,
                                 grouping =cell_names,
                                 title='ebpmf',
                                 print_plot = F,
                                 loadings_order=loadings_order)
  grid.arrange(plot1, plot0,plot2, nrow=3)
}

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

fit = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc_fasttopics/init_tol_effect/nonnegLF_pe_inittol1e6_iter60.rds')
for(i in 1:4){
  plot(fit$sigma2,res$output[[i]]$fitted_model$ebpmf$sigma2,col='grey80',pch=20,xlab='true',ylab='estimated',main='variance estimation')
  abline(a=0,b=1)
}

Large true variance, small init variance

res = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc3k_simulation/simu_pbmc_fasttopics_large_true_var_small_init_var.rds')
K_hat = c()
for(i in 1:length(res$output)){
  K_hat = rbind(K_hat,c(res$output[[i]]$fitted_model$flash$n.factors,res$output[[i]]$fitted_model$ebpmf$fit_flash$n.factors))
}
colnames(K_hat) = c('flash','ebpmf')
K_hat

     flash ebpmf
[1,]     4     9
[2,]     4     9
[3,]     4     9
[4,]     4     8
[5,]     4     9

loadings_order = get_loadings_order(res$sim_data$Loading,res$sim_data$Factor,grouping = cell_names,n_samples = 5000)

Perplexity automatically changed to 53 because the original setting of 100 was too large for the number of samples (163)

plot0=structure_plot_general(res$sim_data$Loading,res$sim_data$Factor,grouping =cell_names,title = 'True',print_plot = F,loadings_order = loadings_order)

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

for(i in 1:length(res$output)){
  plot1 = structure_plot_general(res$output[[i]]$fitted_model$flash$L.pm,
                                 res$output[[i]]$fitted_model$flash$F.pm
                                 ,grouping=cell_names,
                                 title='flash',
                                 print_plot = F,
                                 loadings_order=loadings_order,
                                 remove_l0f0 = F)
  plot2 = structure_plot_general(res$output[[i]]$fitted_model$ebpmf$fit_flash$L.pm,
                                 res$output[[i]]$fitted_model$ebpmf$fit_flash$F.pm,
                                 grouping =cell_names,
                                 title='ebpmf',
                                 print_plot = F,
                                 loadings_order=loadings_order)
  grid.arrange(plot1, plot0,plot2, nrow=3)
}

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

fit = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc_fasttopics/init_tol_effect/nonnegLF_pe_inittol1e2_iter60.rds')
for(i in 1:length(res$output)){
  plot(fit$sigma2,res$output[[i]]$fitted_model$ebpmf$sigma2,col='grey80',pch=20,xlab='true',ylab='estimated',main='variance estimation')
  abline(a=0,b=1)
}

Large true variance, large init variance

res = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc3k_simulation/simu_pbmc_fasttopics_large_true_var_large_init_var.rds')
K_hat = c()
for(i in c(1,3,4)){
  K_hat = rbind(K_hat,c(res$output[[i]]$fitted_model$flash$n.factors,res$output[[i]]$fitted_model$ebpmf$fit_flash$n.factors))
}
colnames(K_hat) = c('flash','ebpmf')
K_hat

     flash ebpmf
[1,]     4    10
[2,]     4    10
[3,]     4    10

loadings_order = get_loadings_order(res$sim_data$Loading,res$sim_data$Factor,grouping = cell_names,n_samples = 5000)

Perplexity automatically changed to 53 because the original setting of 100 was too large for the number of samples (163)

plot0=structure_plot_general(res$sim_data$Loading,res$sim_data$Factor,grouping =cell_names,title = 'True',print_plot = F,loadings_order = loadings_order)

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

for(i in c(1,3,4)){
  plot1 = structure_plot_general(res$output[[i]]$fitted_model$flash$L.pm,
                                 res$output[[i]]$fitted_model$flash$F.pm
                                 ,grouping=cell_names,
                                 title='flash',
                                 print_plot = F,
                                 loadings_order=loadings_order,
                                 remove_l0f0 = F)
  plot2 = structure_plot_general(res$output[[i]]$fitted_model$ebpmf$fit_flash$L.pm,
                                 res$output[[i]]$fitted_model$ebpmf$fit_flash$F.pm,
                                 grouping =cell_names,
                                 title='ebpmf',
                                 print_plot = F,
                                 loadings_order=loadings_order)
  grid.arrange(plot1, plot0,plot2, nrow=3)
}

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

Warning in structure_plot(fit_list, grouping = grouping, loadings_order =
loadings_order, : Input argument "n" is ignored when "loadings_order" is not
"embed"

fit = readRDS('/project2/mstephens/dongyue/poisson_mf/pbmc_fasttopics/init_tol_effect/nonnegLF_pe_inittol1e2_iter60.rds')
for(i in c(1,3,4,5)){
  plot(fit$sigma2,res$output[[i]]$fitted_model$ebpmf$sigma2,col='grey80',pch=20,xlab='true',ylab='estimated',main='variance estimation')
  abline(a=0,b=1)
}

sessionInfo()

R version 4.1.0 (2021-05-18)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS:   /software/R-4.1.0-no-openblas-el7-x86_64/lib64/R/lib/libRblas.so
LAPACK: /software/R-4.1.0-no-openblas-el7-x86_64/lib64/R/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C         LC_TIME=C           
 [4] LC_COLLATE=C         LC_MONETARY=C        LC_MESSAGES=C       
 [7] LC_PAPER=C           LC_NAME=C            LC_ADDRESS=C        
[10] LC_TELEPHONE=C       LC_MEASUREMENT=C     LC_IDENTIFICATION=C 

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] ggplot2_3.4.0      gridExtra_2.3      Matrix_1.5-3       fastTopics_0.6-142
[5] workflowr_1.6.2   

loaded via a namespace (and not attached):
  [1] mcmc_0.9-7         bitops_1.0-7       matrixStats_0.59.0
  [4] fs_1.5.0           progress_1.2.2     httr_1.4.4        
  [7] rprojroot_2.0.2    tools_4.1.0        bslib_0.2.5.1     
 [10] utf8_1.2.2         R6_2.5.1           irlba_2.3.5.1     
 [13] uwot_0.1.14        DBI_1.1.1          lazyeval_0.2.2    
 [16] colorspace_2.0-3   wavethresh_4.7.2   withr_2.5.0       
 [19] tidyselect_1.2.0   prettyunits_1.1.1  ebpm_0.0.1.3      
 [22] compiler_4.1.0     git2r_0.28.0       cli_3.5.0         
 [25] quantreg_5.94      SparseM_1.81       plotly_4.10.1     
 [28] labeling_0.4.2     horseshoe_0.2.0    sass_0.4.0        
 [31] smashrgen_1.1.4    caTools_1.18.2     flashier_0.2.34   
 [34] scales_1.2.1       SQUAREM_2021.1     quadprog_1.5-8    
 [37] pbapply_1.6-0      mixsqp_0.3-48      stringr_1.4.0     
 [40] digest_0.6.30      rmarkdown_2.9      MCMCpack_1.6-3    
 [43] deconvolveR_1.2-1  vebpm_0.4.4        pkgconfig_2.0.3   
 [46] htmltools_0.5.3    ebpmf_2.0.8        highr_0.9         
 [49] fastmap_1.1.0      invgamma_1.1       htmlwidgets_1.5.4 
 [52] rlang_1.0.6        rstudioapi_0.13    farver_2.1.1      
 [55] jquerylib_0.1.4    generics_0.1.3     jsonlite_1.8.3    
 [58] dplyr_1.0.10       magrittr_2.0.3     smashr_1.3-6      
 [61] Rcpp_1.0.9         munsell_0.5.0      fansi_1.0.3       
 [64] RcppZiggurat_0.1.6 lifecycle_1.0.3    stringi_1.6.2     
 [67] whisker_0.4        yaml_2.3.6         MASS_7.3-54       
 [70] plyr_1.8.6         Rtsne_0.16         grid_4.1.0        
 [73] parallel_4.1.0     promises_1.2.0.1   ggrepel_0.9.2     
 [76] crayon_1.5.2       lattice_0.20-44    cowplot_1.1.1     
 [79] splines_4.1.0      hms_1.1.2          knitr_1.33        
 [82] pillar_1.8.1       softImpute_1.4-1   reshape2_1.4.4    
 [85] glue_1.6.2         evaluate_0.14      trust_0.1-8       
 [88] data.table_1.14.6  RcppParallel_5.1.5 nloptr_1.2.2.2    
 [91] vctrs_0.5.1        httpuv_1.6.1       MatrixModels_0.5-1
 [94] gtable_0.3.1       purrr_0.3.5        ebnm_1.0-11       
 [97] tidyr_1.2.1        assertthat_0.2.1   ashr_2.2-54       
[100] xfun_0.24          Rfast_2.0.6        NNLM_0.4.4        
[103] coda_0.19-4        later_1.3.0        survival_3.2-11   
[106] viridisLite_0.4.1  truncnorm_1.0-8    tibble_3.1.8      
[109] ellipsis_0.3.2