Skip to content

Latest commit

 

History

History
1050 lines (853 loc) · 41.9 KB

README.md

File metadata and controls

1050 lines (853 loc) · 41.9 KB

sccomp - Tests differences in cell type proportions and variability from single-cell data

Lifecycle:maturing R build status

Cellular omics such as single-cell genomics, proteomics, and microbiomics allow the characterization of tissue and microbial community composition, which can be compared between conditions to identify biological drivers. This strategy has been critical to unveiling markers of disease progression in conditions such as cancer and pathogen infections.

For cellular omic data, no method for differential variability analysis exists, and methods for differential composition analysis only take a few fundamental data properties into account. Here we introduce sccomp, a generalised method for differential composition and variability analyses capable of jointly modelling data count distribution, compositionality, group-specific variability, and proportion mean-variability association, while being robust to outliers.

sccomp is an extensive analysis framework that allows realistic data simulation and cross-study knowledge transfer. We demonstrate that mean-variability association is ubiquitous across technologies, highlighting the inadequacy of the very popular Dirichlet-multinomial modeling and providing essential principles for differential variability analysis.

Watch the video

Cite

Mangiola, Stefano, Alexandra J. Roth-Schulze, Marie Trussart, Enrique Zozaya-Valdés, Mengyao Ma, Zijie Gao, Alan F. Rubin, Terence P. Speed, Heejung Shim, and Anthony T. Papenfuss. 2023. “Sccomp: Robust Differential Composition and Variability Analysis for Single-Cell Data.” Proceedings of the National Academy of Sciences of the United States of America 120 (33): e2203828120. https://doi.org/10.1073/pnas.2203828120 PNAS - sccomp: Robust differential composition and variability analysis for single-cell data

sccomp tests differences in cell type proportions from single-cell data. It is robust against outliers, it models continuous and discrete factors, and capable of random-effect/intercept modelling.

Characteristics

  • Complex linear models with continuous and categorical covariates
  • Multilevel modelling, with population fixed and random effects/intercept
  • Modelling data from counts
  • Testing differences in cell-type proportionality
  • Testing differences in cell-type specific variability
  • Cell-type information share for variability adaptive shrinkage
  • Testing differential variability
  • Probabilistic outlier identification
  • Cross-dataset learning (hyperpriors).

Installation

sccomp is based on cmdstanr which provides the latest version of cmdstan the Bayesian modelling tool. cmdstanr is not on CRAN, so we need to have 3 simple step process (that will be prompted to the user is forgot).

  1. R installation of sccomp
  2. R installation of cmdstanr
  3. cmdstanr call to cmdstan installation

Bioconductor

if (!requireNamespace("BiocManager")) install.packages("BiocManager")

# Step 1
BiocManager::install("sccomp")

# Step 2
install.packages("cmdstanr", repos = c("https://stan-dev.r-universe.dev/", getOption("repos")))

# Step 3
cmdstanr::check_cmdstan_toolchain(fix = TRUE) # Just checking system setting
cmdstanr::install_cmdstan()

Github

# Step 1
devtools::install_github("MangiolaLaboratory/sccomp")

# Step 2
install.packages("cmdstanr", repos = c("https://stan-dev.r-universe.dev/", getOption("repos")))

# Step 3
cmdstanr::check_cmdstan_toolchain(fix = TRUE) # Just checking system setting
cmdstanr::install_cmdstan()
Function Description
sccomp_estimate Fit the model onto the data, and estimate the coefficients
sccomp_remove_outliers Identify outliers probabilistically based on the model fit, and exclude them from the estimation
sccomp_test Calculate the probability that the coefficients are outside the H0 interval (i.e. test_composition_above_logit_fold_change)
sccomp_replicate Simulate data from the model, or part of the model
sccomp_predict Predicts proportions, based on the model, or part of the model
sccomp_remove_unwanted_variation Removes the variability for unwanted factors
plot Plots summary plots to asses significance

Analysis

library(dplyr)
library(sccomp)
library(ggplot2)
library(forcats)
library(tidyr)
data("seurat_obj")
data("sce_obj")
data("counts_obj")

sccomp can model changes in composition and variability. By default, the formula for variability is either ~1, which assumes that the cell-group variability is independent of any covariate or ~ factor_of_interest, which assumes that the model is dependent on the factor of interest only. The variability model must be a subset of the model for composition.

Binary factor

Of the output table, the estimate columns start with the prefix c_ indicate composition, or with v_ indicate variability (when formula_variability is set).

From Seurat, SingleCellExperiment, metadata objects

sccomp_result = 
  sce_obj |>
  sccomp_estimate( 
    formula_composition = ~ type, 
    .sample =  sample, 
    .cell_group = cell_group, 
    cores = 1 
  ) |> 
  sccomp_remove_outliers(cores = 1) |> # Optional
  sccomp_test()

From counts

sccomp_result = 
  counts_obj |>
  sccomp_estimate( 
    formula_composition = ~ type, 
    .sample = sample,
    .cell_group = cell_group,
    .count = count, 
    cores = 1, verbose = FALSE
  ) |> 
  sccomp_remove_outliers(cores = 1, verbose = FALSE) |> # Optional
  sccomp_test()
## Running standalone generated quantities after 1 MCMC chain, with 1 thread(s) per chain...
## 
## Chain 1 finished in 0.0 seconds.

## Running standalone generated quantities after 1 MCMC chain, with 1 thread(s) per chain...
## 
## Chain 1 finished in 0.0 seconds.

Here you see the results of the fit, the effects of the factor on composition and variability. You also can see the uncertainty around those effects.

The output is a tibble containing the Following columns

  • cell_group - The cell groups being tested.
  • parameter - The parameter being estimated from the design matrix described by the input formula_composition and formula_variability.
  • factor - The covariate factor in the formula, if applicable (e.g., not present for Intercept or contrasts).
  • c_lower - Lower (2.5%) quantile of the posterior distribution for a composition (c) parameter.
  • c_effect - Mean of the posterior distribution for a composition (c) parameter.
  • c_upper - Upper (97.5%) quantile of the posterior distribution for a composition (c) parameter.
  • c_pH0 - Probability of the null hypothesis (no difference) for a composition (c). This is not a p-value.
  • c_FDR - False-discovery rate of the null hypothesis for a composition (c).
  • v_lower - Lower (2.5%) quantile of the posterior distribution for a variability (v) parameter.
  • v_effect - Mean of the posterior distribution for a variability (v) parameter.
  • v_upper - Upper (97.5%) quantile of the posterior distribution for a variability (v) parameter.
  • v_pH0 - Probability of the null hypothesis for a variability (v).
  • v_FDR - False-discovery rate of the null hypothesis for a variability (v).
  • count_data - Nested input count data.
sccomp_result
## # A tibble: 72 × 14
##    cell_group parameter  factor c_lower c_effect c_upper   c_pH0   c_FDR v_lower
##    <chr>      <chr>      <chr>    <dbl>    <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1 B1         (Intercep… <NA>    0.981     1.13   1.28   0       0         -6.33
##  2 B1         typecancer type   -1.15     -0.892 -0.646  0       0         NA   
##  3 B2         (Intercep… <NA>    0.476     0.781  1.08   0       0         -5.64
##  4 B2         typecancer type   -1.20     -0.779 -0.346  5.00e-4 4.55e-5   NA   
##  5 B3         (Intercep… <NA>   -0.614    -0.326 -0.0416 6.27e-2 5.89e-3   -6.83
##  6 B3         typecancer type   -0.622    -0.219  0.196  2.83e-1 7.30e-2   NA   
##  7 BM         (Intercep… <NA>   -1.25     -0.948 -0.667  0       0         -7.44
##  8 BM         typecancer type   -0.738    -0.340  0.0937 1.37e-1 4.42e-2   NA   
##  9 CD4 1      (Intercep… <NA>    0.208     0.385  0.562  2.25e-3 1.30e-4   -6.75
## 10 CD4 1      typecancer type   -0.0828    0.150  0.388  3.38e-1 9.76e-2   NA   
## # ℹ 62 more rows
## # ℹ 5 more variables: v_effect <dbl>, v_upper <dbl>, v_pH0 <dbl>, v_FDR <dbl>,
## #   count_data <list>

Summary plots

A plot of group proportions, faceted by groups. The blue boxplots represent the posterior predictive check. If the model is descriptively adequate for the data, the blue boxplots should roughly overlay the black boxplots, which represent the observed data. The outliers are coloured in red. A boxplot will be returned for every (discrete) covariate present in formula_composition. The colour coding represents the significant associations for composition and/or variability.

sccomp_result |> 
  sccomp_boxplot(factor = "type")
## Loading model from cache...

## Running standalone generated quantities after 1 MCMC chain, with 1 thread(s) per chain...
## 
## Chain 1 finished in 0.0 seconds.

## Joining with `by = join_by(cell_group, sample)`

## Joining with `by = join_by(cell_group, type)`

A plot of estimates of differential composition (c_) on the x-axis and differential variability (v_) on the y-axis. The error bars represent 95% credible intervals. The dashed lines represent the minimal effect that the hypothesis test is based on. An effect is labelled as significant if it exceeds the minimal effect according to the 95% credible interval. Facets represent the covariates in the model.

sccomp_result |> 
  plot_1D_intervals()

We can plot the relationship between abundance and variability. As we can see below, they are positively correlated. sccomp models this relationship to obtain a shrinkage effect on the estimates of both the abundance and the variability. This shrinkage is adaptive as it is modelled jointly, thanks to Bayesian inference.

sccomp_result |> 
  plot_2D_intervals()

You can produce the series of plots calling the plot method.

sccomp_result |> plot() 

Model proportions directly (e.g. from deconvolution)

Note: If counts are available, we strongly discourage the use of proportions, as an important source of uncertainty (i.e., for rare groups/cell types) is not modeled.

The use of proportions is better suited for modelling deconvolution results (e.g., of bulk RNA data), in which case counts are not available.

Proportions should be greater than 0. Assuming that zeros derive from a precision threshold (e.g., deconvolution), zeros are converted to the smallest non-zero value.

Continuous factor

sccomp is able to fit erbitrary complex models. In this example we have a continuous and binary covariate.

res =
    seurat_obj |>
    sccomp_estimate(
      formula_composition = ~ type + continuous_covariate, 
      .sample = sample, .cell_group = cell_group,
      cores = 1, verbose=FALSE
    )
## Loading required package: SeuratObject

## Loading required package: sp

## 
## Attaching package: 'SeuratObject'

## The following objects are masked from 'package:base':
## 
##     intersect, t

## sccomp says: count column is an integer. The sum-constrained beta binomial model will be used

## sccomp says: estimation

## sccomp says: the composition design matrix has columns: (Intercept), typehealthy, continuous_covariate

## sccomp says: the variability design matrix has columns: (Intercept)

## Loading model from cache...

## sccomp says: to do hypothesis testing run `sccomp_test()`,
##   the `test_composition_above_logit_fold_change` = 0.1 equates to a change of ~10%, and
##   0.7 equates to ~100% increase, if the baseline is ~0.1 proportion.
##   Use `sccomp_proportional_fold_change` to convert c_effect (linear) to proportion difference (non-linear).
res
## # A tibble: 90 × 10
##    cell_group parameter factor c_lower c_effect c_upper v_lower v_effect v_upper
##    <chr>      <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>    <dbl>   <dbl>
##  1 B immature (Interce… <NA>    0.377    0.759   1.12     -4.07    -3.67   -3.29
##  2 B immature typeheal… type    0.866    1.36    1.86     NA       NA      NA   
##  3 B immature continuo… conti… -0.250    0.0610  0.378    NA       NA      NA   
##  4 B mem      (Interce… <NA>   -1.25    -0.804  -0.365    -4.93    -4.52   -4.13
##  5 B mem      typeheal… type    1.03     1.67    2.29     NA       NA      NA   
##  6 B mem      continuo… conti… -0.240    0.0896  0.415    NA       NA      NA   
##  7 CD4 cm S1… (Interce… <NA>    1.17     1.49    1.82     -3.66    -3.27   -2.89
##  8 CD4 cm S1… typeheal… type    0.689    1.12    1.56     NA       NA      NA   
##  9 CD4 cm S1… continuo… conti… -0.0662   0.187   0.441    NA       NA      NA   
## 10 CD4 cm hi… (Interce… <NA>   -0.940   -0.460   0.0176   -5.13    -4.63   -4.15
## # ℹ 80 more rows
## # ℹ 1 more variable: count_data <list>

Random Effect Modeling

sccomp supports multilevel modeling by allowing the inclusion of random effects in the compositional and variability formulas. This is particularly useful when your data has hierarchical or grouped structures, such as measurements nested within subjects, batches, or experimental units. By incorporating random effects, sccomp can account for variability at different levels of your data, improving model fit and inference accuracy.

Random Intercept Model

In this example, we demonstrate how to fit a random intercept model using sccomp. We’ll model the cell-type proportions with both fixed effects (e.g., treatment) and random effects (e.g., subject-specific variability).

Here is the input data

seurat_obj[[]] |> as_tibble()
## # A tibble: 106,297 × 9
##    cell_group nCount_RNA nFeature_RNA group__ group__wrong sample type  group2__
##    <chr>           <dbl>        <int> <chr>   <chr>        <chr>  <chr> <chr>   
##  1 CD4 naive           0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  2 Mono clas…          0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  3 CD4 cm S1…          0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  4 B immature          0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  5 CD8 naive           0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  6 CD4 naive           0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  7 Mono clas…          0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  8 CD4 cm S1…          0            0 GROUP1  1            SI-GA… canc… GROUP21 
##  9 CD4 cm hi…          0            0 GROUP1  1            SI-GA… canc… GROUP21 
## 10 B immature          0            0 GROUP1  1            SI-GA… canc… GROUP21 
## # ℹ 106,287 more rows
## # ℹ 1 more variable: continuous_covariate <dbl>
res = 
  seurat_obj |>
  sccomp_estimate( 
    formula_composition = ~ type + (1 | group__), 
    .sample = sample,
    .cell_group = cell_group,
    bimodal_mean_variability_association = TRUE,
    cores = 1, verbose = FALSE
  ) 
## sccomp says: count column is an integer. The sum-constrained beta binomial model will be used

## sccomp says: estimation

## sccomp says: the composition design matrix has columns: (Intercept), typehealthy

## sccomp says: the variability design matrix has columns: (Intercept)

## Loading model from cache...

## sccomp says: to do hypothesis testing run `sccomp_test()`,
##   the `test_composition_above_logit_fold_change` = 0.1 equates to a change of ~10%, and
##   0.7 equates to ~100% increase, if the baseline is ~0.1 proportion.
##   Use `sccomp_proportional_fold_change` to convert c_effect (linear) to proportion difference (non-linear).
res
## # A tibble: 180 × 10
##    cell_group parameter factor c_lower c_effect c_upper v_lower v_effect v_upper
##    <chr>      <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>    <dbl>   <dbl>
##  1 B immature (Interce… <NA>    0.498    0.839   1.17     -4.09    -3.66   -3.14
##  2 B immature typeheal… type    0.687    1.08    1.53     NA       NA      NA   
##  3 B immature (Interce… <NA>   -0.324    0.0821  0.672    NA       NA      NA   
##  4 B immature (Interce… <NA>   -0.0712   0.276   0.796    NA       NA      NA   
##  5 B immature (Interce… <NA>   -0.183    0.204   0.636    NA       NA      NA   
##  6 B immature (Interce… <NA>   -0.744   -0.311   0.0402   NA       NA      NA   
##  7 B mem      (Interce… <NA>   -0.769   -0.281   0.133    -4.96    -4.44   -3.88
##  8 B mem      typeheal… type    0.309    0.959   1.63     NA       NA      NA   
##  9 B mem      (Interce… <NA>   -0.294    0.0881  0.679    NA       NA      NA   
## 10 B mem      (Interce… <NA>   -0.0411   0.335   0.907    NA       NA      NA   
## # ℹ 170 more rows
## # ℹ 1 more variable: count_data <list>

Random Effect Model (random slopes)

sccomp can model random slopes. We providean example below.

res = 
  seurat_obj |>
  sccomp_estimate(
      formula_composition = ~ type + (type | group__),
      .sample = sample,
      .cell_group = cell_group,
      bimodal_mean_variability_association = TRUE,
      cores = 1, verbose = FALSE
    )
## sccomp says: count column is an integer. The sum-constrained beta binomial model will be used

## sccomp says: estimation

## sccomp says: the composition design matrix has columns: (Intercept), typehealthy

## sccomp says: the variability design matrix has columns: (Intercept)

## Loading model from cache...

## sccomp says: to do hypothesis testing run `sccomp_test()`,
##   the `test_composition_above_logit_fold_change` = 0.1 equates to a change of ~10%, and
##   0.7 equates to ~100% increase, if the baseline is ~0.1 proportion.
##   Use `sccomp_proportional_fold_change` to convert c_effect (linear) to proportion difference (non-linear).
res
## # A tibble: 240 × 10
##    cell_group parameter factor c_lower c_effect c_upper v_lower v_effect v_upper
##    <chr>      <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>    <dbl>   <dbl>
##  1 B immature (Interce… <NA>     0.489   0.848   1.29     -4.17    -3.67   -2.97
##  2 B immature typeheal… type     0.490   1.03    1.51     NA       NA      NA   
##  3 B immature (Interce… <NA>    -0.186   0.0634  0.523    NA       NA      NA   
##  4 B immature typeheal… <NA>    -0.215   0.0533  0.494    NA       NA      NA   
##  5 B immature (Interce… <NA>    -0.195   0.142   0.607    NA       NA      NA   
##  6 B immature typeheal… <NA>    -0.159   0.144   0.582    NA       NA      NA   
##  7 B immature (Interce… <NA>    -0.148   0.176   0.611    NA       NA      NA   
##  8 B immature (Interce… <NA>    -0.752  -0.269   0.0267   NA       NA      NA   
##  9 B mem      (Interce… <NA>    -0.738  -0.331   0.0703   -4.97    -4.42   -3.78
## 10 B mem      typeheal… type     0.317   0.923   1.58     NA       NA      NA   
## # ℹ 230 more rows
## # ℹ 1 more variable: count_data <list>

Nested Random Effects

If you have a more complex hierarchy, such as measurements nested within subjects and subjects nested within batches, you can include multiple grouping variables. Here group2__ is nested within group__.

res = 
  seurat_obj |>
  sccomp_estimate(
      formula_composition = ~ type + (type | group__) + (1 | group2__),
      .sample = sample,
      .cell_group = cell_group,
      bimodal_mean_variability_association = TRUE,
      cores = 1, verbose = FALSE
    )
## sccomp says: count column is an integer. The sum-constrained beta binomial model will be used

## sccomp says: estimation

## sccomp says: the composition design matrix has columns: (Intercept), typehealthy

## sccomp says: the variability design matrix has columns: (Intercept)

## Loading model from cache...

## sccomp says: to do hypothesis testing run `sccomp_test()`,
##   the `test_composition_above_logit_fold_change` = 0.1 equates to a change of ~10%, and
##   0.7 equates to ~100% increase, if the baseline is ~0.1 proportion.
##   Use `sccomp_proportional_fold_change` to convert c_effect (linear) to proportion difference (non-linear).
res
## # A tibble: 300 × 10
##    cell_group parameter factor c_lower c_effect c_upper v_lower v_effect v_upper
##    <chr>      <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>    <dbl>   <dbl>
##  1 B immature (Interce… <NA>    0.437    0.820   1.27     -4.28    -3.63   -2.94
##  2 B immature typeheal… type    0.700    1.13    1.67     NA       NA      NA   
##  3 B immature (Interce… <NA>   -0.202    0.0312  0.550    NA       NA      NA   
##  4 B immature typeheal… <NA>   -0.158    0.0252  0.446    NA       NA      NA   
##  5 B immature (Interce… <NA>   -0.192    0.0614  0.365    NA       NA      NA   
##  6 B immature typeheal… <NA>   -0.130    0.0555  0.349    NA       NA      NA   
##  7 B immature (Interce… <NA>   -0.106    0.105   0.534    NA       NA      NA   
##  8 B immature (Interce… <NA>   -0.712   -0.150   0.0377   NA       NA      NA   
##  9 B immature (Interce… <NA>   -0.407   -0.0769  0.0888   NA       NA      NA   
## 10 B immature (Interce… <NA>   -0.0716   0.133   0.609    NA       NA      NA   
## # ℹ 290 more rows
## # ℹ 1 more variable: count_data <list>

An aid to result interpretation and communication

The estimated effects are expressed in the unconstrained space of the parameters, similar to differential expression analysis that expresses changes in terms of log fold change. However, for differences in proportion, logit fold change must be used, which is harder to interpret and understand.

Therefore, we provide a more intuitive proportional fold change that can be more easily understood. However, these cannot be used to infer significance (use sccomp_test() instead), and a lot of care must be taken given the nonlinearity of these measures (a 1-fold increase from 0.0001 to 0.0002 carries a different weight than a 1-fold increase from 0.4 to 0.8).

From your estimates, you can specify which effects you are interested in (this can be a subset of the full model if you wish to exclude unwanted effects), and the two points you would like to compare.

In the case of a categorical variable, the starting and ending points are categories.

sccomp_result |> 
   sccomp_proportional_fold_change(
     formula_composition = ~  type,
     from =  "healthy", 
     to = "cancer"
    ) |> 
  select(cell_group, statement)
## Loading model from cache...

## Running standalone generated quantities after 1 MCMC chain, with 1 thread(s) per chain...
## 
## Chain 1 finished in 0.0 seconds.

## # A tibble: 36 × 2
##    cell_group statement                                
##    <chr>      <glue>                                   
##  1 B1         2.4-fold decrease (from 0.0541 to 0.0223)
##  2 B2         2.2-fold decrease (from 0.0382 to 0.0177)
##  3 B3         1.3-fold decrease (from 0.0129 to 0.0102)
##  4 BM         1.4-fold decrease (from 0.0068 to 0.0049)
##  5 CD4 1      1.1-fold increase (from 0.0258 to 0.0296)
##  6 CD4 2      1.7-fold increase (from 0.0498 to 0.0828)
##  7 CD4 3      3.4-fold decrease (from 0.1111 to 0.0327)
##  8 CD4 4      1.2-fold increase (from 0.0017 to 0.002) 
##  9 CD4 5      1.1-fold increase (from 0.0305 to 0.0325)
## 10 CD8 1      1.3-fold increase (from 0.1 to 0.1254)   
## # ℹ 26 more rows

Contrasts

seurat_obj |>
  sccomp_estimate( 
    formula_composition = ~ 0 + type, 
    .sample = sample,
    .cell_group = cell_group, 
    cores = 1, verbose = FALSE
  ) |> 
  sccomp_test( contrasts =  c("typecancer - typehealthy", "typehealthy - typecancer"))
## # A tibble: 60 × 14
##    cell_group  parameter factor c_lower c_effect c_upper   c_pH0   c_FDR v_lower
##    <chr>       <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1 B immature  typecanc… <NA>    -1.89    -1.35   -0.803 0       0            NA
##  2 B immature  typeheal… <NA>     0.803    1.35    1.89  0       0            NA
##  3 B mem       typecanc… <NA>    -2.26    -1.64   -1.04  0       0            NA
##  4 B mem       typeheal… <NA>     1.04     1.64    2.26  0       0            NA
##  5 CD4 cm S10… typecanc… <NA>    -1.43    -0.993  -0.531 0       0            NA
##  6 CD4 cm S10… typeheal… <NA>     0.531    0.993   1.43  0       0            NA
##  7 CD4 cm hig… typecanc… <NA>     0.847    1.56    2.24  0       0            NA
##  8 CD4 cm hig… typeheal… <NA>    -2.24    -1.56   -0.847 0       0            NA
##  9 CD4 cm rib… typecanc… <NA>     0.264    0.931   1.57  0.00750 0.00180      NA
## 10 CD4 cm rib… typeheal… <NA>    -1.57    -0.931  -0.264 0.00750 0.00180      NA
## # ℹ 50 more rows
## # ℹ 5 more variables: v_effect <dbl>, v_upper <dbl>, v_pH0 <dbl>, v_FDR <dbl>,
## #   count_data <list>

Categorical factor (e.g. Bayesian ANOVA)

This is achieved through model comparison with loo. In the following example, the model with association with factors better fits the data compared to the baseline model with no factor association. For comparisons check_outliers must be set to FALSE as the leave-one-out must work with the same amount of data, while outlier elimination does not guarantee it.

If elpd_diff is away from zero of > 5 se_diff difference of 5, we are confident that a model is better than the other reference. In this case, -79.9 / 11.5 = -6.9, therefore we can conclude that model one, the one with factor association, is better than model two.

library(loo)

# Fit first model
model_with_factor_association = 
  seurat_obj |>
  sccomp_estimate( 
    formula_composition = ~ type, 
    .sample =  sample, 
    .cell_group = cell_group, 
    inference_method = "hmc",
    enable_loo = TRUE
  )
## Running MCMC with 6 parallel chains, with 3 thread(s) per chain...
## 
## Chain 1 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 1 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 1 Iteration: 200 / 966 [ 20%]  (Warmup)

## Chain 2 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 2 Iteration: 100 / 966 [ 10%]  (Warmup)

## Chain 3 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 3 Iteration: 100 / 966 [ 10%]  (Warmup)

## Chain 4 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 4 Iteration: 100 / 966 [ 10%]  (Warmup)

## Chain 5 Iteration:   1 / 966 [  0%]  (Warmup)

## Chain 6 Iteration:   1 / 966 [  0%]  (Warmup)

## Chain 1 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 1 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 2 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 3 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 1 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 2 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 2 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 4 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 5 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 6 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 3 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 3 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 2 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 4 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 4 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 5 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 1 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 3 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 6 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 4 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 5 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 5 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 1 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 2 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 6 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 6 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 3 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 4 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 5 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 1 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 2 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 6 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 3 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 5 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 2 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 4 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 1 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 3 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 6 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 2 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 4 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 5 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 1 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 3 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 6 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 2 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 4 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 5 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 1 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 3 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 6 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 1 finished in 2.3 seconds.
## Chain 2 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 4 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 5 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 2 finished in 2.3 seconds.
## Chain 3 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 4 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 3 finished in 2.3 seconds.
## Chain 4 finished in 2.3 seconds.
## Chain 5 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 6 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 5 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 5 finished in 2.4 seconds.
## Chain 6 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 6 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 6 finished in 2.5 seconds.
## 
## All 6 chains finished successfully.
## Mean chain execution time: 2.4 seconds.
## Total execution time: 3.2 seconds.
# Fit second model
model_without_association = 
  seurat_obj |>
  sccomp_estimate( 
    formula_composition = ~ 1, 
    .sample =  sample, 
    .cell_group = cell_group, 
    inference_method = "hmc",
    enable_loo = TRUE
  )
## Running MCMC with 6 parallel chains, with 3 thread(s) per chain...
## 
## Chain 1 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 1 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 1 Iteration: 200 / 966 [ 20%]  (Warmup)

## Chain 2 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 2 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 2 Iteration: 200 / 966 [ 20%]  (Warmup)

## Chain 3 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 3 Iteration: 100 / 966 [ 10%]  (Warmup)

## Chain 4 Iteration:   1 / 966 [  0%]  (Warmup) 
## Chain 4 Iteration: 100 / 966 [ 10%]  (Warmup)

## Chain 5 Iteration:   1 / 966 [  0%]  (Warmup)

## Chain 6 Iteration:   1 / 966 [  0%]  (Warmup)

## Chain 1 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 1 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 3 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 5 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 2 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 2 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 4 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 6 Iteration: 100 / 966 [ 10%]  (Warmup) 
## Chain 1 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 3 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 3 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 5 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 2 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 4 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 4 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 6 Iteration: 200 / 966 [ 20%]  (Warmup) 
## Chain 1 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 3 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 5 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 5 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 2 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 4 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 6 Iteration: 300 / 966 [ 31%]  (Warmup) 
## Chain 6 Iteration: 301 / 966 [ 31%]  (Sampling) 
## Chain 3 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 5 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 1 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 2 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 4 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 6 Iteration: 400 / 966 [ 41%]  (Sampling) 
## Chain 3 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 1 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 4 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 5 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 2 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 6 Iteration: 500 / 966 [ 51%]  (Sampling) 
## Chain 1 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 3 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 4 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 5 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 2 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 6 Iteration: 600 / 966 [ 62%]  (Sampling) 
## Chain 3 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 4 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 5 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 1 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 2 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 6 Iteration: 700 / 966 [ 72%]  (Sampling) 
## Chain 1 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 3 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 4 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 5 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 1 finished in 2.4 seconds.
## Chain 2 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 3 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 6 Iteration: 800 / 966 [ 82%]  (Sampling) 
## Chain 2 finished in 2.3 seconds.
## Chain 3 finished in 2.3 seconds.
## Chain 4 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 5 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 4 finished in 2.2 seconds.
## Chain 6 Iteration: 900 / 966 [ 93%]  (Sampling) 
## Chain 5 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 5 finished in 2.3 seconds.
## Chain 6 Iteration: 966 / 966 [100%]  (Sampling) 
## Chain 6 finished in 2.3 seconds.
## 
## All 6 chains finished successfully.
## Mean chain execution time: 2.3 seconds.
## Total execution time: 3.0 seconds.
# Compare models
loo_compare(
   attr(model_with_factor_association, "fit")$loo(),
   attr(model_without_association, "fit")$loo()
)
##        elpd_diff se_diff
## model1   0.0       0.0  
## model2 -79.5      10.7

Differential variability, binary factor

We can model the cell-group variability also dependent on the type, and so test differences in variability

res = 
  seurat_obj |>
  sccomp_estimate( 
    formula_composition = ~ type, 
    formula_variability = ~ type,
    .sample = sample,
    .cell_group = cell_group,
    cores = 1, verbose = FALSE
  )

res
## # A tibble: 60 × 10
##    cell_group         parameter factor c_lower c_effect c_upper v_lower v_effect
##    <chr>              <chr>     <chr>    <dbl>    <dbl>   <dbl>   <dbl>    <dbl>
##  1 B immature         (Interce… <NA>     0.330    0.748  1.15    -4.36    -3.95 
##  2 B immature         typeheal… type     0.815    1.36   1.89    -0.871   -0.224
##  3 B mem              (Interce… <NA>    -1.34    -0.864 -0.368   -5.14    -4.65 
##  4 B mem              typeheal… type     1.08     1.72   2.36    -1.41    -0.683
##  5 CD4 cm S100A4      (Interce… <NA>     1.33     1.67   2.01    -3.79    -3.40 
##  6 CD4 cm S100A4      typeheal… type     0.387    0.835  1.28    -1.19    -0.687
##  7 CD4 cm high cytok… (Interce… <NA>    -1.03    -0.511  0.0200  -5.13    -4.62 
##  8 CD4 cm high cytok… typeheal… type    -1.91    -1.08  -0.144    0.477    1.31 
##  9 CD4 cm ribosome    (Interce… <NA>    -0.152    0.330  0.816   -4.90    -4.38 
## 10 CD4 cm ribosome    typeheal… type    -1.74    -1.06  -0.393   -0.357    0.261
## # ℹ 50 more rows
## # ℹ 2 more variables: v_upper <dbl>, count_data <list>

Plot 1D significance plot

plots = res |> sccomp_test() |> plot()
## Loading model from cache...

## Running standalone generated quantities after 1 MCMC chain, with 1 thread(s) per chain...
## 
## Chain 1 finished in 0.0 seconds.

## Joining with `by = join_by(cell_group, sample)`

## Joining with `by = join_by(cell_group, type)`
plots$credible_intervals_1D

Plot 2D significance plot Data points are cell groups. Error bars are the 95% credible interval. The dashed lines represent the default threshold fold change for which the probabilities (c_pH0, v_pH0) are calculated. pH0 of 0 represent the rejection of the null hypothesis that no effect is observed.

This plot is provided only if differential variability has been tested. The differential variability estimates are reliable only if the linear association between mean and variability for (intercept) (left-hand side facet) is satisfied. A scatterplot (besides the Intercept) is provided for each category of interest. For each category of interest, the composition and variability effects should be generally uncorrelated.

plots$credible_intervals_2D

Suggested settings

For single-cell RNA sequencing

We recommend setting bimodal_mean_variability_association = TRUE. The bimodality of the mean-variability association can be confirmed from the plots$credible_intervals_2D (see below).

For CyTOF and microbiome data

We recommend setting bimodal_mean_variability_association = FALSE (Default).

Visualisation of the MCMC chains from the posterior distribution

It is possible to directly evaluate the posterior distribution. In this example, we plot the Monte Carlo chain for the slope parameter of the first cell type. We can see that it has converged and is negative with probability 1.

library(cmdstanr)
## This is cmdstanr version 0.8.1

## - CmdStanR documentation and vignettes: mc-stan.org/cmdstanr

## - CmdStan path: /Users/a1234450/.cmdstan/cmdstan-2.35.0

## - CmdStan version: 2.35.0
library(posterior)
## This is posterior version 1.6.0

## 
## Attaching package: 'posterior'

## The following objects are masked from 'package:stats':
## 
##     mad, sd, var

## The following objects are masked from 'package:base':
## 
##     %in%, match
library(bayesplot)
## This is bayesplot version 1.11.1

## - Online documentation and vignettes at mc-stan.org/bayesplot

## - bayesplot theme set to bayesplot::theme_default()

##    * Does _not_ affect other ggplot2 plots

##    * See ?bayesplot_theme_set for details on theme setting

## 
## Attaching package: 'bayesplot'

## The following object is masked from 'package:posterior':
## 
##     rhat
# Assuming res contains the fit object from cmdstanr
fit <- res |> attr("fit")

# Extract draws for 'beta[2,1]'
draws <- as_draws_array(fit$draws("beta[2,1]"))

# Create a traceplot for 'beta[2,1]'
mcmc_trace(draws, pars = "beta[2,1]")

The old framework

The new tidy framework was introduced in 2024, two, understand the differences and improvements. Compared to the old framework, please read this blog post.