pbmc3k 10X result

Last updated: 2023-02-16

Checks: 7 0

Knit directory: gsmash/

This reproducible R Markdown analysis was created with workflowr (version 1.7.0). 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: 8545d58

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 8545d58. 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/

Untracked files:
    Untracked:  analysis/movielens.Rmd
    Untracked:  analysis/profiled_obj_for_b_splitting.Rmd
    Untracked:  code/poisson_STM/structure_plot.R
    Untracked:  data/ml-latest-small/
    Untracked:  data/pbmc3k_10x/
    Untracked:  output/droplet_iteration_results/
    Untracked:  output/ebpmf_pbmc3k_vga3_glmpca_init.rds
    Untracked:  output/pbmc3k_10x/
    Untracked:  output/pbmc3k_iteration_results/
    Untracked:  output/pbmc_no_constraint.rds

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/pbmc3k_10X_result.Rmd) and HTML (docs/pbmc3k_10X_result.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	8545d58	DongyueXie	2023-02-16	wflow_publish("analysis/pbmc3k_10X_result.Rmd")

Data

The pbmc3k data is from 10X, and I downloaded Gene/cell matrix(raw) dataset.

The quality control and cell types annotation are done by Seurat, following the tutorial here.

The resulting dataset has 2638 cells and 13714 genes.

pbmc3k = readRDS('data/pbmc3k_10x/pbmc3k.rds')
dim(pbmc3k$counts)

Loading required package: Matrix

NULL

The cell types are

table(pbmc3k$cell_type)


 Naive CD4 T   CD14+ Mono Memory CD4 T            B        CD8 T FCGR3A+ Mono 
         711          480          472          344          279          162 
          NK           DC     Platelet 
         144           32           14

Filter out genes

I filtered out genes that expressed in less than 10 cells, and no further preprocessing. The final data matrix is

library(stm)
library(Matrix)
counts = pbmc3k$counts
counts_filtered = counts[,colSums(counts!=0)>10]
dim(counts_filtered)

[1]  2638 10908

pbmc3k_sparse = ebpmf_log(counts_filtered,flash_control = list())
saveRDS(pbmc3k_sparse,'pbmc3k_sparse.rds')
pbmc3k_nonnegL = ebpmf_log(counts_filtered,flash_control = list(ebnm.fn=c(ebnm::ebnm_point_exponential,ebnm::ebnm_point_normal),loadings_sign = 1))
saveRDS(pbmc3k_nonnegL,'pbmc3k_nonnegL_pe.rds')
pbmc3k_nonnegLF = ebpmf_log(counts_filtered,flash_control = list(ebnm.fn=c(ebnm::ebnm_point_exponential,ebnm::ebnm_point_exponential),factors_sign=1,loadings_sign = 1))
saveRDS(pbmc3k_nonnegLF,'pbmc3k_nonnegLF_pe.rds')

pbmc3k_nonnegL = ebpmf_log(counts_filtered,flash_control = list(ebnm.fn=c(ebnm::ebnm_unimodal_nonnegative,ebnm::ebnm_point_normal),loadings_sign = 1))
saveRDS(pbmc3k_nonnegL,'pbmc3k_nonnegL_un.rds')
pbmc3k_nonnegLF = ebpmf_log(counts_filtered,flash_control = list(ebnm.fn=c(ebnm::ebnm_unimodal_nonnegative,ebnm::ebnm_unimodal_nonnegative),factors_sign=1,loadings_sign = 1))
saveRDS(pbmc3k_nonnegLF,'pbmc3k_nonnegLF_un.rds')

Model fitting

The model is the Empirical Bayes Poisson matrix factorization model, and \(l_0\) is fixed at log(rowMeans(Y)), \(f_0\) is initialized at log(colSums(Y)/sum(l_0), then let the model estiamtes \(f_0\) during iterations.

summary_plot = function(res){
  plot(res$K_trace,ylab='',xlab='iterations',main='K over iterations',pch=20)
  plot(res$sigma2,ylab='',xlab='gene',main = 'gene specific variance',pch=20,col='grey70')
  plot(colSums(counts_filtered/rowSums(counts_filtered)),res$sigma2,xlab='gene expression',ylab='gene specific varaince',pch=20,col='grey60')
  plot(colSums(counts_filtered==0)/nrow(counts_filtered),res$sigma2,xlab='gene expression sparsity',ylab='gene specific varaince',pch = 20, col='grey50')
}

Sparse loadings and factors

source('code/poisson_STM/plot_factors.R')
pbmc3k_sparse = readRDS('output/pbmc3k_10x/pbmc3k_sparse.rds')

summary_plot(pbmc3k_sparse)

plot.factors(pbmc3k_sparse$fit_flash,pbmc3k$cell_type)

Nonnegative loadings with point exponential prior, and sparse factors

source('code/poisson_STM/structure_plot.R')
pbmc3k_nonnegL_pe = readRDS('output/pbmc3k_10x/pbmc3k_nonnegL_pe.rds')

summary_plot(pbmc3k_nonnegL_pe)

plot.factors(pbmc3k_nonnegL_pe$fit_flash,pbmc3k$cell_type,nonnegative = T)

structure_plot_general(pbmc3k_nonnegL_pe$fit_flash$L.pm,pbmc3k_nonnegL_pe$fit_flash$F.pm,pbmc3k$cell_type,LD=T)

Running tsne on 532 x 7 matrix.

Running tsne on 363 x 7 matrix.

Running tsne on 368 x 7 matrix.

Running tsne on 259 x 7 matrix.

Running tsne on 214 x 7 matrix.

Running tsne on 119 x 7 matrix.

Running tsne on 110 x 7 matrix.

Running tsne on 24 x 7 matrix.

Nonnegative loadings with unimodal nonnegative prior, and sparse factors

pbmc3k_nonnegL_un = readRDS('output/pbmc3k_10x/pbmc3k_nonnegL_un.rds')

summary_plot(pbmc3k_nonnegL_un)

plot.factors(pbmc3k_nonnegL_un$fit_flash,pbmc3k$cell_type,nonnegative = T)

structure_plot_general(pbmc3k_nonnegL_un$fit_flash$L.pm,pbmc3k_nonnegL_un$fit_flash$F.pm,pbmc3k$cell_type,LD=T)

Running tsne on 532 x 6 matrix.

Running tsne on 363 x 6 matrix.

Running tsne on 368 x 6 matrix.

Running tsne on 259 x 6 matrix.

Running tsne on 214 x 6 matrix.

Running tsne on 119 x 6 matrix.

Running tsne on 110 x 6 matrix.

Running tsne on 24 x 6 matrix.

Nonnegative loadings and factors with point exponential prior

I got a warning message saying that

‘Warning message: In scale.EF(EF) : Fitting stopped after the initialization function failed to find a non-zero factor.’

pbmc3k_nonnegLF_pe = readRDS('output/pbmc3k_10x/pbmc3k_nonnegLF_pe.rds')

summary_plot(pbmc3k_nonnegLF_pe)

plot.factors(pbmc3k_nonnegLF_pe$fit_flash,pbmc3k$cell_type,nonnegative = T)

structure_plot_general(pbmc3k_nonnegLF_pe$fit_flash$L.pm,pbmc3k_nonnegLF_pe$fit_flash$F.pm,pbmc3k$cell_type,LD=T)

Running tsne on 532 x 5 matrix.

Running tsne on 363 x 5 matrix.

Running tsne on 368 x 5 matrix.

Running tsne on 259 x 5 matrix.

Running tsne on 214 x 5 matrix.

Running tsne on 119 x 5 matrix.

Running tsne on 110 x 5 matrix.

Running tsne on 24 x 5 matrix.

Nonnegative loadings and factors with unimodal nonnegative prior