--- title: "bettr" author: "Charlotte Soneson & Federico Marini" date: "`r Sys.Date()`" output: BiocStyle::html_document vignette: > %\VignetteIndexEntry{bettr} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ```{r setup, include = FALSE} knitr::opts_chunk$set( collapse = TRUE, comment = "#>" ) ``` # Introduction Method benchmarking is a core part of computational biology research, with an intrinsic power to establish best practices in method selection and application, as well as help identifying gaps and possibilities for improvement. A typical benchmark evaluates a set of methods using multiple different metrics, intended to capture different aspects of their performance. The best method to choose in any given situation can then be found, e.g., by averaging the different performance metrics, possibly putting more emphasis on those that are more important to the specific situation. Inspired by the [OECD 'Better Life Index'](https://www.oecdbetterlifeindex.org/#/11111111111), the `bettr` package was developed to provide support for this last step. It allows users to easily create performance summaries emphasizing the aspects that are most important to them. `bettr` can be used interactively, via a R/shiny application, or programmatically by calling the underlying functions. In this vignette, we illustrate both alternatives, using example data provided with the package. Given the abundance of methods available for computational analysis of biological data, both within and beyond Bioconductor, and the importance of careful, adaptive benchmarking, we believe that `bettr` will be a useful complement to currently available Bioconductor infrastructure related to benchmarking and performance estimation. Other packages (e.g., `r Biocpkg("pipeComp")` or `r Biocpkg("SummarizedBenchmark")`) provide frameworks for _executing_ benchmarks by applying and recording pre-defined workflows to data. Packages such as `r Biocpkg("iCOBRA")` and `r CRANpkg("ROCR")` instead provide functionality for calculating well-established evaluation metric. In contrast, `bettr` focuses on _visual exploration_ of benchmark results, represented by the values of several evaluation metrics. # Installation `bettr` can be installed from Bioconductor (from release 3.19 onwards): ```{r, eval=FALSE} if (!require("BiocManager", quietly = TRUE)) install.packages("BiocManager") BiocManager::install("bettr") ``` # Usage ```{r} suppressPackageStartupMessages({ library("bettr") library("SummarizedExperiment") library("tibble") library("dplyr") }) ``` The main input to `bettr` is a `data.frame` containing values of several metrics for several methods. In addition, the user can provide additional annotations and characteristics for the methods and metrics, which can be used to group and filter them in the interactive application. ```{r} ## Data for two metrics (metric1, metric2) for three methods (M1, M2, M3) df <- data.frame(Method = c("M1", "M2", "M3"), metric1 = c(1, 2, 3), metric2 = c(3, 1, 2)) ## More information for metrics metricInfo <- data.frame(Metric = c("metric1", "metric2", "metric3"), Group = c("G1", "G2", "G2")) ## More information for methods ('IDs') idInfo <- data.frame(Method = c("M1", "M2", "M3"), Type = c("T1", "T1", "T2")) ``` To simplify handling and sharing, the data can be combined into a `SummarizedExperiment` (with methods as rows and metrics as columns) as follows: ```{r} se <- assembleSE(df = df, idCol = "Method", metricInfo = metricInfo, idInfo = idInfo) se ``` The interactive application to explore the rankings can then be launched by means of the `bettr()` function. The input can be either the assembled `SummarizedExperiment` object or the individual components. ```{r} #| eval: false ## Alternative 1 bettr(bettrSE = se) ## Alternative 2 bettr(df = df, idCol = "Method", metricInfo = metricInfo, idInfo = idInfo) ``` # Example - single-cell RNA-seq clustering benchmark Next, we show a more elaborate example, visualizing data from the benchmark of single-cell clustering methods performed by [Duo et al (2018)](https://f1000research.com/articles/7-1141). The values for a set of evaluation metrics applied to results obtained by several clustering methods are provided in a `.csv` file in the package: ```{r} res <- read.csv(system.file("extdata", "duo2018_results.csv", package = "bettr")) dim(res) tibble(res) ``` As we can see, we have 14 methods (rows) and 48 different metrics (columns). The first column provides the name of the clustering method. More precisely, the columns correspond to four different metrics, each of which was applied to clustering output from of 12 data sets. We encode this "grouping" of metrics in a data frame, in such a way that we can later collapse performance across data sets in `bettr`: ```{r} metricInfo <- tibble(Metric = colnames(res)[-1]) |> mutate(Class = sub("_.*", "", Metric)) head(metricInfo) table(metricInfo$Class) ``` In order to make different metrics comparable, we next define the transformation that should be applied to each of them within `bettr`. First, we need to make sure that the metric are consistent in terms of whether large values indicate "good" or "bad" performance. In our case, for both the `elapsed` (elapsed run time), `nclust.vs.true` (difference between estimated and true number of clusters) and `s.norm.vs.true` (difference between estimated and true normalized Shannon entropy for a clustering), a small value indicates "better" performance, while for the `ARI` (adjusted Rand index), larger values are better. Hence, we will flip the sign of the first three before doing additional analyses. Moreover, the different metrics clearly live in different numeric ranges - the maximal value of the `ARI` is 1, while the other metrics can have much larger values. As an example, here we therefore scale the three other metrics linearly to the interval `[0, 1]` to make them more comparable to the `ARI` values. We record these transformations in a list, that will be passed to `bettr`: ```{r} ## Initialize list initialTransforms <- lapply(res[, grep("elapsed|nclust.vs.true|s.norm.vs.true", colnames(res), value = TRUE)], function(i) { list(flip = TRUE, transform = '[0,1]') }) length(initialTransforms) names(initialTransforms) head(initialTransforms) ``` We can specify four different aspects of the desired transform, which will be applied in the following order: * `flip` (`TRUE` or `FALSE`, whether to flip the sign of the values). The default is `FALSE`. * `offset` (a numeric value to add to the observed values, possibly after applying the sign flip). The default is 0. * `transform` (one of `None`, `[0,1]`, `[-1,1]`, `z-score`, or `Rank`). The default is `None`. * `cuts` (a numeric vector of cuts that will be used to turn a numeric variable into a categorical one). The default is `NULL`. Only values that deviate from the defaults need to be specified. Finally, we can define a set of colors that we would like to use for visualizing the methods and metrics in `bettr`. ```{r} metricColors <- list( Class = c(ARI = "purple", elapsed = "forestgreen", nclust.vs.true = "blue", s.norm.vs.true = "orange")) idColors <- list( method = c( CIDR = "#332288", FlowSOM = "#6699CC", PCAHC = "#88CCEE", PCAKmeans = "#44AA99", pcaReduce = "#117733", RtsneKmeans = "#999933", Seurat = "#DDCC77", SC3svm = "#661100", SC3 = "#CC6677", TSCAN = "grey34", ascend = "orange", SAFE = "black", monocle = "red", RaceID2 = "blue" )) ``` All the information defined so far can be combined in a `SummarizedExperiment` object, as shown above for the small example data: ```{r} duo2018 <- assembleSE(df = res, idCol = "method", metricInfo = metricInfo, initialTransforms = initialTransforms, metricColors = metricColors, idColors = idColors) duo2018 ``` The `assay` of the `SummarizedExperiment` object contains the values for the 48 performance measures for the 14 clustering methods. The `metricInfo` is stored in the `colData`, and the lists of colors and the initial transforms in the `metadata`: ```{r} ## Display the whole performance table tibble(assay(duo2018, "values")) ## Showing the first metric, evaluated on all datasets head(colData(duo2018), 12) ## These are the color definitions (can mix character and hex values) metadata(duo2018)$bettrInfo$idColors metadata(duo2018)$bettrInfo$metricColors names(metadata(duo2018)$bettrInfo$initialTransforms) ## An example of a transformation - elapsed time for the Koh dataset metadata(duo2018)$bettrInfo$initialTransforms$elapsed_Koh ``` Now, we can launch the app for this data set: ```{r} #| eval: false bettr(bettrSE = duo2018, bstheme = "sandstone") ``` The screenshot below illustrates the default view of the interactive interface. We can choose to collapse the metric values to have a single value for each metric class, to reduce the redundancy. We can now also freely decide how to weight the respective metrics by means of the sliders in the left side bar. The bars on top of the heatmap show the current weight assignment. `bettr` also provides alternative visualizations, e.g. a polar plot: # Programmatic interface The interactive application showcased above, is the main entry point to using `bettr`. However, we also provide a wrapper function to prepare the input data for plotting (replicating the steps that are performed in the app), as well as access to the plotting functions themselves. The following code replicates the results for the example above. ```{r} #| fig.width: 7 #| fig.height: 7 ## Assign a higher weight to one of the collapsed metric classes metadata(duo2018)$bettrInfo$initialWeights["Class_ARI"] <- 0.55 prepData <- bettrGetReady( bettrSE = duo2018, idCol = "method", scoreMethod = "weighted mean", metricGrouping = "Class", metricCollapseGroup = TRUE) ## This object is fairly verbose and detailed, ## but has the whole set of info needed prepData ## Call the plotting routines specifying one single parameter makeHeatmap(bettrList = prepData) makePolarPlot(bettrList = prepData) ``` # Exporting data from the app It is possible to export the data used internally by the interactive application, in the same format as the output from `bettrGetReady()`. To enable such export, first generate the `app` object using the `bettr()` function, and then assign the call to `shiny::runApp()` to a variable to capture the output. For example: ```{r} if (interactive()) { app <- bettr(bettrSE = duo2018, bstheme = "sandstone") out <- shiny::runApp(app) } ``` To activate the export, make sure to click the button 'Close app' (in the bottom of the left-hand side bar) in order to close the application (don't just close the window). This will take you back to your R session, where the variable `out` will be populated with the data used in the app (in the same format as the output from `bettrGetReady()`). This list can be directly provided as the input to e.g. `makeHeatmap()` and the other plotting functions via the `bettrList` argument, as shown above. # Additional examples `bettr` can also be adapted to represent more types of such collections of metrics, other than the results of a benchmarking study in computational biology. An [example](https://github.com/federicomarini/bettr/blob/devel/inst/scripts/visualize_oecd_data.R), which is also included in the `inst/scripts` folder of this package, presents the OECD Better Life Index (https://stats.oecd.org/index.aspx?DataSetCode=BLI), spanning over 11 topics, each represented by one to three indicators. These indicators are good measures of the concepts of well-being, and well suited to display some comparison across countries. Additional examples can be added to the codebase upon interest, and we encourage users to contribute to that via a Pull Request to https://github.com/federicomarini/bettr. # Session info {-} ```{r} sessionInfo() ```