Installation

To install and load NBAMSeq

Introduction

High-throughput sequencing experiments followed by differential expression analysis is a widely used approach to detect genomic biomarkers. A fundamental step in differential expression analysis is to model the association between gene counts and covariates of interest. NBAMSeq is a flexible statistical model based on the generalized additive model and allows for information sharing across genes in variance estimation. Specifically, we model the logarithm of mean gene counts as sums of smooth functions with the smoothing parameters and coefficients estimated simultaneously by a nested iteration. The variance is estimated by the Bayesian shrinkage approach to fully exploit the information across all genes.

The workflow of NBAMSeq contains three main steps:

Here we illustrate each of these steps respectively.

Data input

Users are expected to provide three parts of input, i.e. countData, colData, and design.

countData is a matrix of gene counts generated by RNASeq experiments.

      sample1 sample2 sample3 sample4 sample5 sample6 sample7 sample8 sample9
gene1       1      87     199       1       4      22    1400     450      53
gene2      11     387     210     323     163     682     139      20       4
gene3       4      47      25      53      23      15     611       3      27
gene4     150     334     113     478      19      27       3      10     164
gene5       4     385      80       1       3       4     218       2     505
gene6     209     294       3      75     408       1      37      32      65
      sample10 sample11 sample12 sample13 sample14 sample15 sample16 sample17
gene1       72       18        2      166      215       15        2      101
gene2        5      502      313       23      374        1       73      106
gene3      160       46      166       10       17      970      920       53
gene4       12      275        1        2      189      286       21      220
gene5       73       23       22       27       15       25      189       34
gene6        1        3      129      311       54        2      195        5
      sample18 sample19 sample20
gene1      255        1       66
gene2        2      300      684
gene3      109      134      800
gene4       24       69       28
gene5       18       36      108
gene6        2      317        3

colData is a data frame which contains the covariates of samples. The sample order in colData should match the sample order in countData.

           pheno       var1        var2       var3 var4
sample1 53.34750 -0.2247691  0.06446786 -0.3313058    0
sample2 72.48195  0.6733607  0.79475848 -1.0492868    2
sample3 43.61273 -0.2916392  0.60957580  0.7930560    2
sample4 75.09485  0.8864967  0.80274262  0.1688156    2
sample5 23.16734  0.7224085  1.14767780  0.1862156    1
sample6 44.15943  0.2235846 -1.33699687  1.1669175    2

design is a formula which specifies how to model the samples. Compared with other packages performing DE analysis including DESeq2 (Love, Huber, and Anders 2014), edgeR (Robinson, McCarthy, and Smyth 2010), NBPSeq (Di et al. 2015) and BBSeq (Zhou, Xia, and Wright 2011), NBAMSeq supports the nonlinear model of covariates via mgcv (Wood and Wood 2015). To indicate the nonlinear covariate in the model, users are expected to use s(variable_name) in the design formula. In our example, if we would like to model pheno as a nonlinear covariate, the design formula should be:

Several notes should be made regarding the design formula:

We then construct the NBAMSeqDataSet using countData, colData, and design:

class: NBAMSeqDataSet 
dim: 50 20 
metadata(1): fitted
assays(1): counts
rownames(50): gene1 gene2 ... gene49 gene50
rowData names(0):
colnames(20): sample1 sample2 ... sample19 sample20
colData names(5): pheno var1 var2 var3 var4

Differential expression analysis

Differential expression analysis can be performed by NBAMSeq function:

Several other arguments in NBAMSeq function are available for users to customize the analysis.

Pulling out DE results

Results of DE analysis can be pulled out by results function. For continuous covariates, the name argument should be specified indicating the covariate of interest. For nonlinear continuous covariates, base mean, effective degrees of freedom (edf), test statistics, p-value, and adjusted p-value will be returned.

DataFrame with 6 rows and 7 columns
       baseMean       edf        stat     pvalue      padj       AIC       BIC
      <numeric> <numeric>   <numeric>  <numeric> <numeric> <numeric> <numeric>
gene1  158.0845   1.00009 1.07928e+01 0.00101972 0.0509859   229.764   236.734
gene2  204.5902   1.00007 1.84654e-02 0.89208222 0.9632527   257.366   264.336
gene3  173.8494   1.00006 8.98823e-01 0.34310322 0.6483236   254.144   261.114
gene4  113.2686   1.00008 5.77412e+00 0.01627357 0.1356131   231.971   238.941
gene5   64.7502   1.00006 2.00604e-01 0.65425958 0.8593302   218.827   225.797
gene6   75.9488   1.00010 1.23881e-05 0.99940303 0.9994030   220.336   227.306

For linear continuous covariates, base mean, estimated coefficient, standard error, test statistics, p-value, and adjusted p-value will be returned.

DataFrame with 6 rows and 8 columns
       baseMean      coef        SE      stat    pvalue      padj       AIC
      <numeric> <numeric> <numeric> <numeric> <numeric> <numeric> <numeric>
gene1  158.0845  0.345618  0.442422  0.781194 0.4346883  0.658619   229.764
gene2  204.5902  0.682681  0.400622  1.704051 0.0883716  0.381821   257.366
gene3  173.8494 -0.582570  0.418137 -1.393250 0.1635443  0.469899   254.144
gene4  113.2686  0.150845  0.368883  0.408924 0.6825956  0.812614   231.971
gene5   64.7502 -0.494415  0.399450 -1.237740 0.2158127  0.469899   218.827
gene6   75.9488  0.572533  0.422086  1.356438 0.1749598  0.469899   220.336
            BIC
      <numeric>
gene1   236.734
gene2   264.336
gene3   261.114
gene4   238.941
gene5   225.797
gene6   227.306

For discrete covariates, the contrast argument should be specified. e.g. contrast = c("var4", "2", "0") means comparing level 2 vs. level 0 in var4.

DataFrame with 6 rows and 8 columns
       baseMean      coef        SE      stat    pvalue      padj       AIC
      <numeric> <numeric> <numeric> <numeric> <numeric> <numeric> <numeric>
gene1  158.0845 -0.623765  0.900235 -0.692891  0.488378  0.775429   229.764
gene2  204.5902  0.424675  0.815115  0.521000  0.602367  0.775429   257.366
gene3  173.8494 -1.162471  0.849030 -1.369176  0.170944  0.470373   254.144
gene4  113.2686 -0.389122  0.751983 -0.517461  0.604834  0.775429   231.971
gene5   64.7502  0.802061  0.809797  0.990447  0.321956  0.703024   218.827
gene6   75.9488 -0.550028  0.861496 -0.638456  0.523177  0.775429   220.336
            BIC
      <numeric>
gene1   236.734
gene2   264.336
gene3   261.114
gene4   238.941
gene5   225.797
gene6   227.306

Visualization

We suggest two approaches to visualize the nonlinear associations. The first approach is to plot the smooth components of a fitted negative binomial additive model by plot.gam function in mgcv (Wood and Wood 2015). This can be done by calling makeplot function and passing in NBAMSeqDataSet object. Users are expected to provide the phenotype of interest in phenoname argument and gene of interest in genename argument.

In addition, to explore the nonlinear association of covariates, it is also instructive to look at log normalized counts vs. variable scatter plot. Below we show how to produce such plot.

DataFrame with 6 rows and 7 columns
        baseMean       edf      stat     pvalue      padj       AIC       BIC
       <numeric> <numeric> <numeric>  <numeric> <numeric> <numeric> <numeric>
gene1   158.0845   1.00009  10.79280 0.00101972 0.0509859   229.764   236.734
gene44  195.7867   1.00007   8.43072 0.00369108 0.0922769   251.798   258.769
gene38   94.7503   1.00007   7.22037 0.00721015 0.1201691   229.573   236.543
gene25  135.2984   1.00011   6.61899 0.01009145 0.1249981   227.938   234.908
gene26   27.3826   1.00009   6.23869 0.01249981 0.1249981   184.299   191.269
gene4   113.2686   1.00008   5.77412 0.01627357 0.1356131   231.971   238.941

Session info

R version 4.0.3 (2020-10-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows Server 2012 R2 x64 (build 9600)

Matrix products: default

locale:
[1] LC_COLLATE=C                          
[2] LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] parallel  stats4    stats     graphics  grDevices utils     datasets 
[8] methods   base     

other attached packages:
 [1] ggplot2_3.3.2               BiocParallel_1.24.0        
 [3] NBAMSeq_1.6.1               SummarizedExperiment_1.20.0
 [5] Biobase_2.50.0              GenomicRanges_1.42.0       
 [7] GenomeInfoDb_1.26.0         IRanges_2.24.0             
 [9] S4Vectors_0.28.0            BiocGenerics_0.36.0        
[11] MatrixGenerics_1.2.0        matrixStats_0.57.0         

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.5             locfit_1.5-9.4         lattice_0.20-41       
 [4] snow_0.4-3             digest_0.6.27          R6_2.5.0              
 [7] RSQLite_2.2.1          evaluate_0.14          httr_1.4.2            
[10] pillar_1.4.6           zlibbioc_1.36.0        rlang_0.4.8           
[13] annotate_1.68.0        blob_1.2.1             Matrix_1.2-18         
[16] rmarkdown_2.5          labeling_0.4.2         splines_4.0.3         
[19] geneplotter_1.68.0     stringr_1.4.0          RCurl_1.98-1.2        
[22] bit_4.0.4              munsell_0.5.0          DelayedArray_0.16.0   
[25] compiler_4.0.3         xfun_0.18              pkgconfig_2.0.3       
[28] mgcv_1.8-33            htmltools_0.5.0        tidyselect_1.1.0      
[31] tibble_3.0.4           GenomeInfoDbData_1.2.4 XML_3.99-0.5          
[34] withr_2.3.0            crayon_1.3.4           dplyr_1.0.2           
[37] bitops_1.0-6           grid_4.0.3             nlme_3.1-150          
[40] xtable_1.8-4           gtable_0.3.0           lifecycle_0.2.0       
[43] DBI_1.1.0              magrittr_1.5           scales_1.1.1          
[46] stringi_1.5.3          farver_2.0.3           XVector_0.30.0        
[49] genefilter_1.72.0      ellipsis_0.3.1         vctrs_0.3.4           
[52] generics_0.0.2         RColorBrewer_1.1-2     tools_4.0.3           
[55] bit64_4.0.5            glue_1.4.2             DESeq2_1.30.0         
[58] purrr_0.3.4            survival_3.2-7         yaml_2.2.1            
[61] AnnotationDbi_1.52.0   colorspace_1.4-1       memoise_1.1.0         
[64] knitr_1.30            

References

Di, Y, DW Schafer, JS Cumbie, and JH Chang. 2015. “NBPSeq: Negative Binomial Models for Rna-Sequencing Data.” R Package Version 0.3. 0, URL Http://CRAN. R-Project. Org/Package= NBPSeq.

Love, Michael I, Wolfgang Huber, and Simon Anders. 2014. “Moderated Estimation of Fold Change and Dispersion for Rna-Seq Data with Deseq2.” Genome Biology 15 (12). BioMed Central:550.

Robinson, Mark D, Davis J McCarthy, and Gordon K Smyth. 2010. “EdgeR: A Bioconductor Package for Differential Expression Analysis of Digital Gene Expression Data.” Bioinformatics 26 (1). Oxford University Press:139–40.

Wood, Simon, and Maintainer Simon Wood. 2015. “Package ’Mgcv’.” R Package Version 1:29.

Zhou, Yi-Hui, Kai Xia, and Fred A Wright. 2011. “A Powerful and Flexible Approach to the Analysis of Rna Sequence Count Data.” Bioinformatics 27 (19). Oxford University Press:2672–8.