bioassayR: Cross-Target Analysis of Small Molecule Bioactivity

Large-Scale Modeling of Sparse Protein Kinase Activity Data

Protein kinases are a protein family that plays an important role in several complex diseases such as cancer and cardiovascular and immunological diseases. Protein kinases have conserved ATP binding sites, which when targeted can lead to similar activities of inhibitors against different kinases. This can be exploited to create multitarget drugs. On the other hand, selectivity (lack of similar activities) is desirable in order to avoid toxicity issues. There is a vast amount of protein kinase activity data in the public domain, which can be used in many different ways. Multitask machine learning models are expected to excel for these kinds of data sets because they can learn from implicit correlations between tasks (in this case activities against a variety of kinases). However, multitask modeling of sparse data poses two major challenges: (i) creating a balanced train-test split without data leakage and (ii) handling missing data. In this work, we construct a protein kinase benchmark set composed of two balanced splits without data leakage, using random and dissimilarity-driven cluster-based mechanisms, respectively. This data set can be used for benchmarking and developing protein kinase activity prediction models. Overall, the performance on the dissimilarity-driven cluster-based split is lower than on random split-based sets for all models, indicating poor generalizability of models. Nevertheless, we show that multitask deep learning models, on this very sparse data set, outperform single-task deep learning and tree-based models. Finally, we demonstrate that data imputation does not improve the performance of (multitask) models on this benchmark set.

Knockdown of the Ribosomal Protein eL38 in HEK293 Cells Changes the Translational Efficiency of Specific Genes

The protein eL38 is one of the smallest proteins of the mammalian ribosome, which is a component of its large (60S) subunit. The haploinsufficiency of eL38 in mice leads to the Tail-short mutant phenotype characterized by defects in the development of the axial skeleton caused by the poor translation of mRNA subsets of Hox genes. Using the ribosome profiling assay applied to HEK293 cells knocked down of eL38, we examined the effects of the lack of eL38 in 60S subunits on gene expression at the level of translation. A four-fold decrease in the cell content of eL38 was shown to result in significant changes in the translational efficiencies of 150 genes. Among the genes, whose expression at the level of translation was enhanced, there were mainly those associated with basic metabolic processes; namely, translation, protein folding, chromosome organization, splicing, and others. The set of genes with reduced translation efficiencies contained those that are mostly involved in the processes related to the regulation of transcription, including the activation of Hox genes. Thus, we demonstrated that eL38 insufficiency significantly affects the expression of certain genes at the translational level. Our findings facilitate understanding the possible causes of some anomalies in eL38-deficient animals.

signatureSearch: environment for gene expression signature searching and functional interpretation

signatureSearch is an R/Bioconductor package that integrates a suite of existing and novel algorithms into an analysis environment for gene expression signature (GES) searching combined with functional enrichment analysis (FEA) and visualization methods to facilitate the interpretation of the search results. In a typical GES search (GESS), a query GES is searched against a database of GESs obtained from large numbers of measurements, such as different genetic backgrounds, disease states and drug perturbations. Database matches sharing correlated signatures with the query indicate related cellular responses frequently governed by connected mechanisms, such as drugs mimicking the expression responses of a disease. To identify which processes are predominantly modulated in the GESS results, we developed specialized FEA methods combined with drug-target network visualization tools. The provided analysis tools are useful for studying the effects of genetic, chemical and environmental perturbations on biological systems, as well as searching single cell GES databases to identify novel network connections or cell types. The signatureSearch software is unique in that it provides access to an integrated environment for GESS/FEA routines that includes several novel search and enrichment methods, efficient data structures, and access to pre-built GES databases, and allowing users to work with custom databases.

High-throughput small molecule screening reveals Nrf2-dependent and -independent pathways of cellular stress resistance

Aging is the dominant risk factor for most chronic diseases. Development of antiaging interventions offers the promise of preventing many such illnesses simultaneously. Cellular stress resistance is an evolutionarily conserved feature of longevity. Here, we identify compounds that induced resistance to the superoxide generator paraquat (PQ), the heavy metal cadmium (Cd), and the DNA alkylator methyl methanesulfonate (MMS). Some rescue compounds conferred resistance to a single stressor, while others provoked multiplex resistance. Induction of stress resistance in fibroblasts was predictive of longevity extension in a published large-scale longevity screen in Caenorhabditis elegans, although not in testing performed in worms and flies with a more restricted set of compounds. Transcriptomic analysis and genetic studies implicated Nrf2/SKN-1 signaling in stress resistance provided by two protective compounds, cardamonin and AEG 3482. Small molecules identified in this work may represent attractive tools to elucidate mechanisms of stress resistance in mammalian cells.

BioMedR: an R/CRAN package for integrated data analysis pipeline in biomedical study

BACKGROUND

With the increasing development of biotechnology and information technology, publicly available data in chemistry and biology are undergoing explosive growth. Such wealthy information in these resources needs to be extracted and then transformed to useful knowledge by various data mining methods. However, a main computational challenge is how to effectively represent or encode molecular objects under investigation such as chemicals, proteins, DNAs and even complicated interactions when data mining methods are employed. To further explore these complicated data, an integrated toolkit to represent different types of molecular objects and support various data mining algorithms is urgently needed.

RESULTS

We developed a freely available R/CRAN package, called BioMedR, for molecular representations of chemicals, proteins, DNAs and pairwise samples of their interactions. The current version of BioMedR could calculate 293 molecular descriptors and 13 kinds of molecular fingerprints for small molecules, 9920 protein descriptors based on protein sequences and six types of generalized scale-based descriptors for proteochemometric modeling, more than 6000 DNA descriptors from nucleotide sequences and six types of interaction descriptors using three different combining strategies. Moreover, this package realized five similarity calculation methods and four powerful clustering algorithms as well as several useful auxiliary tools, which aims at building an integrated analysis pipeline for data acquisition, data checking, descriptor calculation and data modeling.

CONCLUSION

BioMedR provides a comprehensive and uniform R package to link up different representations of molecular objects with each other and will benefit cheminformatics/bioinformatics and other biomedical users. It is available at: https://CRAN.R-project.org/package=BioMedR and https://github.com/wind22zhu/BioMedR/.

CSgator: an integrated web platform for compound set analysis

Drug discovery typically involves investigation of a set of compounds (e.g. drug screening hits) in terms of target, disease, and bioactivity. CSgator is a comprehensive analytic tool for set-wise interpretation of compounds. It has two unique analytic features of Compound Set Enrichment Analysis (CSEA) and Compound Cluster Analysis (CCA), which allows batch analysis of compound set in terms of (i) target, (ii) bioactivity, (iii) disease, and (iv) structure. CSEA and CCA present enriched profiles of targets and bioactivities in a compound set, which leads to novel insights on underlying drug mode-of-action, and potential targets. Notably, we propose a novel concept of 'Hit Enriched Assays", i.e. bioassays of which hits are enriched among a given set of compounds. As an example, we show its utility in revealing drug mode-of-action or identifying hidden targets for anti-lymphangiogenesis screening hits. CSgator is available at http://csgator.ewha.ac.kr , and most analytic results are downloadable.

Effect of missing data on multitask prediction methods

There has been a growing interest in multitask prediction in chemoinformatics, helped by the increasing use of deep neural networks in this field. This technique is applied to multitarget data sets, where compounds have been tested against different targets, with the aim of developing models to predict a profile of biological activities for a given compound. However, multitarget data sets tend to be sparse; i.e., not all compound-target combinations have experimental values. There has been little research on the effect of missing data on the performance of multitask methods. We have used two complete data sets to simulate sparseness by removing data from the training set. Different models to remove the data were compared. These sparse sets were used to train two different multitask methods, deep neural networks and Macau, which is a Bayesian probabilistic matrix factorization technique. Results from both methods were remarkably similar and showed that the performance decrease because of missing data is at first small before accelerating after large amounts of data are removed. This work provides a first approximation to assess how much data is required to produce good performance in multitask prediction exercises.

Large-scale bioactivity analysis of the small-molecule assayed proteome

This study presents an analysis of the small molecule bioactivity profiles across large quantities of diverse protein families represented in PubChem BioAssay. We compared the bioactivity profiles of FDA approved drugs to non-FDA approved compounds, and report several distinct patterns characteristic of the approved drugs. We found that a large fraction of the previously reported higher target promiscuity among FDA approved compounds, compared to non-FDA approved bioactives, was frequently due to cross-reactivity within rather than across protein families. We identified 804 potentially novel protein target candidates for FDA approved drugs, as well as 901 potentially novel target candidates with active non-FDA approved compounds, but no FDA approved drugs with activity against these targets. We also identified 486348 potentially novel compounds active against the same targets as FDA approved drugs, as well as 153402 potentially novel compounds active against targets without active FDA approved drugs. By quantifying the agreement among replicated screens, we estimated that more than half of these novel outcomes are reproducible. Using biclustering, we identified many dense clusters of FDA approved drugs with enriched activity against a common set of protein targets. We also report the distribution of compound promiscuity using a Bayesian statistical model, and report the sensitivity and specificity of two common methods for identifying promiscuous compounds. Aggregator assays exhibited greater accuracy in identifying highly promiscuous compounds, while PAINS substructures were able to identify a much larger set of "middle range" promiscuous compounds. Additionally, we report a large number of promiscuous compounds not identified as aggregators or PAINS. In summary, the results of this study represent a rich reference for selecting novel drug and target protein candidates, as well as for eliminating candidate compounds with unselective activities.