Complex models are commonly used in predictive modeling. In this paper we present R packages that can be used for explaining predictions from complex black box models and attributing parts of these predictions to input features. We introduce two new approaches and corresponding packages for such attribution, namely live and breakDown. We also compare their results with existing implementations of state-of-the-art solutions, namely lime that implements Locally Interpretable Model-agnostic Explanations and ShapleyR that implements Shapley values.

The MSstats R-Bioconductor family of packages is widely used for statistical analyses of quantitative bottom-up mass spectrometry-based proteomic experiments to detect differentially abundant proteins. It is applicable to a variety of experimental designs and data acquisition strategies and is compatible with many data processing tools used to identify and quantify spectral features. In the face of ever-increasing complexities of experiments and data processing strategies, the core package of the family, with the same name MSstats, has undergone a series of substantial updates. Its new version MSstats v4.0 improves the usability, versatility, and accuracy of statistical methodology, and the usage of computational resources. New converters integrate the output of upstream processing tools directly with MSstats, requiring less manual work by the user. The package’s statistical models have been updated to a more robust workflow. Finally, MSstats’ code has been substantially refactored to improve memory use and computation speed. Here we detail these updates, highlighting methodological differences between the new and old versions. An empirical comparison of MSstats v4.0 to its previous implementations, as well as to the packages MSqRob and DEqMS, on controlled mixtures and biological experiments demonstrated a stronger performance and better usability of MSstats v4.0 as compared to existing methods.

PurposeThis paper seeks to develop universal software (a programming framework) enabling the implementation of service robot controllers. The software should distinguish the hardware‐oriented part of the system from the task‐oriented one. Moreover, force, vision as well as other sensors should be taken into account. Multi‐effector systems have to be considered.Design/methodology/approachThe robot programming framework MRROC++ has been implemented as a hierarchical structure composed of processes, potentially consisting of threads. All of the software is written in an object‐oriented manner using C++ and is supervised by a QNX real‐time operating system. The framework has been verified on several systems executing diverse tasks. Here, a Rubik's cube puzzle‐solving system, consisting of two arms and utilizing force control and visual servos, is presented.FindingsThe presented framework is well suited to tasks requiring two‐handed manipulation with force sensing, visual servoing and online construction of plans of actions. The Rubik's cube puzzle is a reasonable initial benchmark for validation of fundamental service robot capabilities. It requires force sensing and sight coupled with two‐handed manipulation and logical reasoning, as do the majority of service tasks. Owing to the use of force sensing during manipulation, jamming of the faces has always been avoided; however, visual servoing could only cope with slow handing over of the cube due to the volume of computations associated with vision processing.Research limitations/implicationsThe proposed software structure does not limit the implementation of service robot controllers. However, some of the specific algorithms used for the solution of the benchmark task (i.e. Rubik's cube puzzle) need to be less time‐consuming.Practical implicationsThe MRROC++ robot programming framework can be applied to the implementation of diverse robot controllers executing complex service tasks.Originality/valueA demanding benchmark task for service robots has been formulated. This task, as well as many others, has been used to validate the MRROC++ robot programming framework which significantly facilitates the implementation of diverse robot systems.

No abstract

Liquid chromatography coupled with bottom-up mass spectrometry (LC-MS/ MS)-based proteomics is a versatile technology for identifying and quantifying proteins in complex biological mixtures. Postidentification, analysis of changes in protein abundances between conditions requires increasingly complex and specialized statistical methods. Many of these methods, in particular the family of open-source Bioconductor packages MSstats, are implemented in a coding language such as R. To make the methods in MSstats accessible to users with limited programming and statistical background, we have created MSstatsShiny, an R-Shiny graphical user interface (GUI) integrated with MSstats, MSstatsTMT, and MSstatsPTM. The GUI provides a point and click analysis pipeline applicable to a wide variety of proteomics experimental types, including label-free data-dependent acquisitions (DDAs) or data-independent acquisitions (DIAs), or tandem mass tag (TMT)-based TMT-DDAs, answering questions such as relative changes in the abundance of peptides, proteins, or post-translational modifications (PTMs). To support reproducible research, the application saves user's selections and builds an R script that programmatically recreates the analysis. MSstatsShiny can be installed locally via Github and Bioconductor, or utilized on the cloud at www.msstatsshiny.com. We illustrate the utility of the platform using two experimental data sets (MassIVE IDs MSV000086623 and MSV000085565).