Skip to contents

Benchmarking HTRfit and DESeq2

04-htrfit_vs_deseq2.Rmd

HTRfit offers a wrapper for DESeq2 outputs. This functionality allows users to seamlessly integrate the results obtained from DESeq2 into the HTRfit evaluation pipeline. By doing so, you can readily compare the performance of HTRfit with DESeq2 on your RNAseq data. This comparative analysis aids in determining which tool performs better for your specific research goals and dataset

Simulation

The choice of input parameters plays a critical role in the simulation process. For optimal results, we recommend basing these decisions on real data, as outlined in the Simulation tutorial

## -- init a design 
list_var <- init_variable( name = "genotype", sd = 0.2462256, level = 2) %>%
            init_variable( name = "environment", sd = 0.2462256, level = 4) %>%
            add_interaction( between_var = c("genotype", "environment"), sd = 0.2462256)
              

N_GENES <- 100
MIN_REPLICATES <- 4
MAX_REPLICATES <- 4
SEQ_DEPTH <- 5e6

## -- simulate RNAseq data 
mock_data <- mock_rnaseq(list_var, 
                         n_genes = N_GENES,
                         min_replicates  = MIN_REPLICATES,
                         max_replicates = MAX_REPLICATES,
                         basal_expression = 2,
                         sequencing_depth = SEQ_DEPTH)

Fit models 

## -- data from simulation
count_matrix <- mock_data$counts
metaData <- mock_data$metadata

HTRfit 

## -- convert counts matrix and samples metadatas in a data frame for fitting
data2fit = prepareData2fit(countMatrix = count_matrix, 
                           metadata =  metaData, 
                           normalization = 'MRN',
                           response_name = "kij")

l_tmb <- fitModelParallel(
          formula = kij ~ genotype + environment  + genotype:environment,
          data = data2fit, 
          group_by = "geneID",
          family = glmmTMB::nbinom2(link = "log"), 
          n.cores = 1)

DESeq2 

## -- DESeq2
dds <- DESeq2::DESeqDataSetFromMatrix(
  countData = count_matrix,
  colData = metaData,
  design = ~ genotype + environment  + genotype:environment )
dds <- DESeq2::DESeq(dds, quiet = TRUE)

Evaluation

Using evaluation_report() function, we assess the capability of DESeq2 and HTRfit to identify conditions exhibiting a 25% change in expression. Based on the identity plot, ROC curve, and PR curve, we observe similar results between the two methods.

## -- get simulation/fit evaluation
resSimu <- evaluation_report(list_tmb = l_tmb,
                             dds = dds,
                             mock_obj = mock_data,
                             coeff_threshold = log(1.25), ## 25% of expression change
                             alt_hypothesis = "greaterAbs")
## -- Model params
resSimu$identity$params

## -- dispersion 
resSimu$identity$dispersion

## -- precision-recall curve
resSimu$precision_recall$params

## -- ROC curve
resSimu$roc$params

Finally, we summarize the model’s performance of HTRfit and DESeq2 using various metrics, including R-squared (R²) and root mean square error (RMSE) for regression tasks, as well as area under the curve (AUC) for classification tasks.

## -- actual/estimate comparison
idx <- c(3,4,5,8,9,10)
resSimu$performances$byparams[idx, c('from', 'description',  'R2', 'RMSE')]
from description R2 RMSE
DESeq2 environment 0.7663712 0.2120142
DESeq2 genotype 0.9805118 0.0566179
DESeq2 genotype:environment 0.7791885 0.1344705
HTRfit environment 0.7663018 0.2120535
HTRfit genotype 0.9804845 0.0566553
HTRfit genotype:environment 0.7791163 0.1344988 
## -- classification metrics
idx <- c(3,4,5,8,9,10)
resSimu$performances$byparams[idx, c('from', 'description', 'pr_AUC',   
                                     'pr_randm_AUC', 'pr_performance_ratio',
                                     'roc_AUC', 'roc_randm_AUC')]
from description pr_AUC pr_randm_AUC pr_performance_ratio roc_AUC roc_randm_AUC
DESeq2 environment 0.9730370 0.5633333 1.727285 0.9628484 0.5
DESeq2 genotype 0.9794961 0.5000000 1.958992 0.9732000 0.5
DESeq2 genotype:environment 0.8945609 0.3466667 2.580464 0.9311960 0.5
HTRfit environment 0.9654006 0.5633333 1.713729 0.9382538 0.5
HTRfit genotype 0.9762610 0.5000000 1.952522 0.9612000 0.5
HTRfit genotype:environment 0.9153040 0.3466667 2.640300 0.9203051 0.5