Designing agricultural grasses to help mitigate proteolysis during ensiling to optimize protein feed provisions for livestock.

The efficient preservation of protein in silage for livestock feed is dependent on the rate and extent of proteolysis. Previous research on fresh forage indicated enhanced protein stability in certain Festulolium (ryegrass × fescue hybrids) cultivars compared to ryegrass. This is the first report of an experiment to test the hypothesis that a Lolium perenne × Festuca arundinacea var glaucescens cultivar had reduced proteolysis compared to perennial ryegrass (L. perenne) during the ensiling process. Forages were harvested in May (Cut 2) and August (Cut 4), wilted for 24 h and ensiled in laboratory-scale silos. Silage was destructively sampled at 0 h, 9 h, 24 h, 48 h, 72 h, 14 days and 90 days post-ensiling, and dry matter (DM), pH and chemical composition were determined. At Cut 2, there was no difference in crude protein between treatments but ryegrass had higher soluble nitrogen (SN) (P < 0.001) and grass × time interactions (p = 0.03) indicated higher rates of proteolysis. By Cut 4, Festulolium had (5.5% units) higher CP than ryegrass (p < 0.001) but SN did not differ. Ammonia-N did not differ between silages in either cut. DM differences (11.8% units) between treatments in Cut 4 (v.2.2% in Cut 2) may have masked effects on proteolysis, highlighting the importance of management on silage quality. This was despite higher WSC in ryegrass in both cuts (p < 0.001), with grass × time interactions (Cut 2; p = 0.03) showing slower WSC decline in ryegrass in Cut 4 (p < 0.001). Silage pH values did not differ between grasses in either cut, but grass × time interactions (p < 0.001) showed a slower decline in both ryegrass cuts, resulting in higher (p < 0.05) pH at 24 h and 72 h for Cuts 2 and 4, respectively. Overall, the hypothesis for an enhanced protein stability in Festulolium when ensiled as ruminant feed was evidenced by lower SN but not ammonia-N in an early-cut silage with a comparable DM to ryegrass.

Do agricultural grasses bred for improved root systems provide resilience to machinery-derived soil compaction?

The increasing frequency of droughts and floods on grasslands, due to climate change, increases the risk of soil compaction. Soil compaction affects both soil and forage productivity. Differing grasses may counteract some effects of compaction due to differences in their root architecture and ontogeny. To compare their resilience to soil compaction, three Festulolium (ryegrass and fescue species' hybrids) forage grass cultivars comprising differing root architecture and ontogeny were compared in replicated field plots, together with a ryegrass and tall fescue variety as controls. Pre-compaction soil and forage properties were determined in spring using > four-year-old plots to generate baseline data. Half of each field plot was then artificially compacted using farm machinery. Forage dry matter yield (DMY) was determined over four cuts. After the final harvest, post compaction soil characteristics and root biomass (RB) were compared between grasses in the non-compacted and compacted soils. Pre-compaction data showed that soil under Festulolium and ryegrass had similar water infiltration rates, higher than soil under tall fescue plots. Tiller density of the Festulolium at this time was significantly higher than fescue but not the ryegrass control. Forage DMY was significantly lower (p < .001) with compacted soil at the first cut but, by the completion of the growing season, there was no effect of soil compaction on total DMY. Tall fescue had a higher total DMY than other grasses, which all produced similar annual yields. Soil bulk density and penetration resistance were higher, and grass tiller density was lower in compacted soils. Root biomass in compacted soils showed a tendency for Festulolium cv Lp × Fg to have higher RB than the ryegrass at 0-15 cm depth. Overall, findings showed alternative grass root structures provide differing resilience to machinery compaction, and root biomass production can be encouraged without negative impacts on forage productivity.

A multi-scale convolutional neural network for phenotyping high-content cellular images.

MOTIVATION: Identifying phenotypes based on high-content cellular images is challenging. Conventional image analysis pipelines for phenotype identification comprise multiple independent steps, with each step requiring method customization and adjustment of multiple parameters. RESULTS: Here, we present an approach based on a multi-scale convolutional neural network (M-CNN) that classifies, in a single cohesive step, cellular images into phenotypes by using directly and solely the images' pixel intensity values. The only parameters in the approach are the weights of the neural network, which are automatically optimized based on training images. The approach requires no a priori knowledge or manual customization, and is applicable to single- or multi-channel images displaying single or multiple cells. We evaluated the classification performance of the approach on eight diverse benchmark datasets. The approach yielded overall a higher classification accuracy compared with state-of-the-art results, including those of other deep CNN architectures. In addition to using the network to simply obtain a yes-or-no prediction for a given phenotype, we use the probability outputs calculated by the network to quantitatively describe the phenotypes. This study shows that these probability values correlate with chemical treatment concentrations. This finding validates further our approach and enables chemical treatment potency estimation via CNNs. AVAILABILITY AND IMPLEMENTATION: The network specifications and solver definitions are provided in Supplementary Software 1. CONTACT: william_jose.godinez_navarro@novartis.com or xian-1.zhang@novartis.com. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Dark chemical matter as a promising starting point for drug lead discovery.

High-throughput screening (HTS) is an integral part of early drug discovery. Herein, we focused on those small molecules in a screening collection that have never shown biological activity despite having been exhaustively tested in HTS assays. These compounds are referred to as 'dark chemical matter' (DCM). We quantified DCM, validated it in quality control experiments, described its physicochemical properties and mapped it into chemical space. Through analysis of prospective reporter-gene assay, gene expression and yeast chemogenomics experiments, we evaluated the potential of DCM to show biological activity in future screens. We demonstrated that, despite the apparent lack of activity, occasionally these compounds can result in potent hits with unique activity and clean safety profiles, which makes them valuable starting points for lead optimization efforts. Among the identified DCM hits was a new antifungal chemotype with strong activity against the pathogen Cryptococcus neoformans but little activity at targets relevant to human safety.

The opportunities of mining historical and collective data in drug discovery.

Vast amounts of bioactivity data have been generated for small molecules across public and corporate domains. Biological signatures, either derived from systematic profiling efforts or from existing historical assay data, have been successfully employed for small molecule mechanism-of-action elucidation, drug repositioning, hit expansion and screening subset design. This article reviews different types of biological descriptors and applications, and we demonstrate how biological data can outlive the original purpose or project for which it was generated. By comparing 150 HTS campaigns run at Novartis over the past decade on the basis of their active and inactive chemical matter, we highlight the opportunities and challenges associated with cross-project learning in drug discovery.

Effects of chicory/perennial ryegrass swards compared with perennial ryegrass swards on the performance and carcass quality of grazing beef steers.

An experiment investigated whether the inclusion of chicory (Cichorium intybus) in swards grazed by beef steers altered their performance, carcass characteristics or parasitism when compared to steers grazing perennial ryegrass (Lolium perenne). Triplicate 2-ha plots were established with a chicory/ryegrass mix or ryegrass control. Forty-eight Belgian Blue-cross steers were used in the first grazing season and a core group (n = 36) were retained for finishing in the second grazing season. The experiment comprised of a standardisation and measurement period. During standardisation, steers grazed a ryegrass/white clover pasture as one group. Animals were allocated to treatment on the basis of liveweight, body condition and faecal egg counts (FEC) determined 7 days prior to the measurement period. The measurement period ran from 25 May until 28 September 2010 and 12 April until 11 October 2011 in the first and second grazing year. Steers were weighed every 14 days at pasture or 28 days during housing. In the first grazing year, faecal samples were collected for FEC and parasite cultures. At the end of the first grazing year, individual blood samples were taken to determine O. ostertagi antibody and plasma pepsinogen levels. During winter, animals were housed as one group and fed silage. In the second grazing year, steers were slaughtered when deemed to reach fat class 3. Data on steer performance showed no differences in daily live-weight gain which averaged 1.04 kg/day. The conformation, fat grade and killing out proportion of beef steers grazing chicory/ryegrass or ryegrass were not found to differ. No differences in FEC, O. ostertagi antibody or plasma pepsinogen levels of beef steers grazing either chicory/ryegrass or ryegrass were observed. Overall, there were no detrimental effects of including chicory in swards grazed by beef cattle on their performance, carcass characteristics or helminth parasitism, when compared with steers grazing ryegrass.

Rethinking molecular similarity: comparing compounds on the basis of biological activity.

Since the advent of high-throughput screening (HTS), there has been an urgent need for methods that facilitate the interrogation of large-scale chemical biology data to build a mode of action (MoA) hypothesis. This can be done either prior to the HTS by subset design of compounds with known MoA or post HTS by data annotation and mining. To enable this process, we developed a tool that compares compounds solely on the basis of their bioactivity: the chemical biological descriptor "high-throughput screening fingerprint" (HTS-FP). In the current embodiment, data are aggregated from 195 biochemical and cell-based assays developed at Novartis and can be used to identify bioactivity relationships among the in-house collection comprising ~1.5 million compounds. We demonstrate the value of the HTS-FP for virtual screening and in particular scaffold hopping. HTS-FP outperforms state of the art methods in several aspects, retrieving bioactive compounds with remarkable chemical dissimilarity to a probe structure. We also apply HTS-FP for the design of screening subsets in HTS. Using retrospective data, we show that a biodiverse selection of plates performs significantly better than a chemically diverse selection of plates, both in terms of number of hits and diversity of chemotypes retrieved. This is also true in the case of hit expansion predictions using HTS-FP similarity. Sets of compounds clustered with HTS-FP are biologically meaningful, in the sense that these clusters enrich for genes and gene ontology (GO) terms, showing that compounds that are bioactively similar also tend to target proteins that operate together in the cell. HTS-FP are valuable not only because of their predictive power but mainly because they relate compounds solely on the basis of bioactivity, harnessing the accumulated knowledge of a high-throughput screening facility toward the understanding of how compounds interact with the proteome.

Chemotography for multi-target SAR analysis in the context of biological pathways.

The increasing amount of chemogenomics data, that is, activity measurements of many compounds across a variety of biological targets, allows for better understanding of pharmacology in a broad biological context. Rather than assessing activity at individual biological targets, today understanding of compound interaction with complex biological systems and molecular pathways is often sought in phenotypic screens. This perspective poses novel challenges to structure-activity relationship (SAR) assessment. Today, the bottleneck of drug discovery lies in the understanding of SAR of rich datasets that go beyond single targets in the context of biological pathways, potential off-targets, and complex selectivity profiles. To aid in the understanding and interpretation of such complex SAR, we introduce Chemotography (chemotype chromatography), which encodes chemical space using a color spectrum by combining clustering and multidimensional scaling. Rich biological data in our approach were visualized using spatial dimensions traditionally reserved for chemical space. This allowed us to analyze SAR in the context of target hierarchies and phylogenetic trees, two-target activity scatter plots, and biological pathways. Chemotography, in combination with the Kyoto Encyclopedia of Genes and Genomes (KEGG), also allowed us to extract pathway-relevant SAR from the ChEMBL database. We identified chemotypes showing polypharmacology and selectivity-conferring scaffolds, even in cases where individual compounds have not been tested against all relevant targets. In addition, we analyzed SAR in ChEMBL across the entire Kinome, going beyond individual compounds. Our method combines the strengths of chemical space visualization for SAR analysis and graphical representation of complex biological data. Chemotography is a new paradigm for chemogenomic data visualization and its versatile applications presented here may allow for improved assessment of SAR in biological context, such as phenotypic assay hit lists.

A lead discovery strategy driven by a comprehensive analysis of proteases in the peptide substrate space.

We present here a comprehensive analysis of proteases in the peptide substrate space and demonstrate its applicability for lead discovery. Aligned octapeptide substrates of 498 proteases taken from the MEROPS peptidase database were used for the in silico analysis. A multiple-category naïve Bayes model, trained on the two-dimensional chemical features of the substrates, was able to classify the substrates of 365 (73%) proteases and elucidate statistically significant chemical features for each of their specific substrate positions. The positional awareness of the method allows us to identify the most similar substrate positions between proteases. Our analysis reveals that proteases from different families, based on the traditional classification (aspartic, cysteine, serine, and metallo), could have substrates that differ at the cleavage site (P1-P1') but are similar away from it. Caspase-3 (cysteine protease) and granzyme B (serine protease) are previously known examples of cross-family neighbors identified by this method. To assess whether peptide substrate similarity between unrelated proteases could reliably translate into the discovery of low molecular weight synthetic inhibitors, a lead discovery strategy was tested on two other cross-family neighbors--namely cathepsin L2 and matrix metallo proteinase 9, and calpain 1 and pepsin A. For both these pairs, a naïve Bayes classifier model trained on inhibitors of one protease could successfully enrich those of its neighbor from a different family and vice versa, indicating that this approach could be prospectively applied to lead discovery for a novel protease target with no known synthetic inhibitors.

Plate-based diversity selection based on empirical HTS data to enhance the number of hits and their chemical diversity.

Typically, screening collections of pharmaceutical companies contain more than a million compounds today. However, for certain high-throughput screening (HTS) campaigns, constraints posed by the assay throughput and/or the reagent costs make it impractical to screen the entire deck. Therefore, it is desirable to effectively screen subsets of the collection based on a hypothesis or a diversity selection. How to select compound subsets is a subject of ongoing debate. The authors present an approach based on extended connectivity fingerprints to carry out diversity selection on a per plate basis (instead of a per compound basis). HTS data from 35 Novartis screens spanning 5 target classes were investigated to assess the performance of this approach. The analysis shows that selecting a fingerprint-diverse subset of 250K compounds, representing 20% of the screening deck, would have achieved significantly higher hit rates for 86% of the screens. This measure also outperforms the Murcko scaffold-based plate selection described previously, where only 49% of the screens showed similar improvements. Strikingly, the 2-fold improvement in average hit rates observed for 3 of 5 target classes in the data set indicates a target bias of the plate (and thus compound) selection method. Even though the diverse subset selection lacks any target hypothesis, its application shows significantly better results for some targets-namely, G-protein-coupled receptors, proteases, and protein-protein interactions-but not for kinase and pathway screens. The synthetic origin of the compounds in the diverse subset appears to influence the screening hit rates. Natural products were the most diverse compound class, with significantly higher hit rates compared to the compounds from the traditional synthetic and combinatorial libraries. These results offer empirical guidelines for plate-based diversity selection to enhance hit rates, based on target class and the library type being screened.

Gaining insight into off-target mediated effects of drug candidates with a comprehensive systems chemical biology analysis.

We present a workflow that leverages data from chemogenomics based target predictions with Systems Biology databases to better understand off-target related toxicities. By analyzing a set of compounds that share a common toxic phenotype and by comparing the pathways they affect with pathways modulated by nontoxic compounds we are able to establish links between pathways and particular adverse effects. We further link these predictive results with literature data in order to explain why a certain pathway is predicted. Specifically, relevant pathways are elucidated for the side effects rhabdomyolysis and hypotension. Prospectively, our approach is valuable not only to better understand toxicities of novel compounds early on but also for drug repurposing exercises to find novel uses for known drugs.

Mapping adverse drug reactions in chemical space.

We present a novel method to better investigate adverse drug reactions in chemical space. By integrating data sources about adverse drug reactions of drugs with an established cheminformatics modeling method, we generate a data set that is then visualized with a systems biology tool. Thereby new insights into undesired drug effects are gained. In this work, we present a global analysis linking chemical features to adverse drug reactions.

Use of ligand based models for protein domains to predict novel molecular targets and applications to triage affinity chromatography data.

The elucidation of drug targets is important both to optimize desired compound action and to understand drug side-effects. In this study, we created statistical models which link chemical substructures of ligands to protein domains in a probabilistic manner and employ the model to triage the results of affinity chromatography experiments. By annotating targets with their InterPro domains, general rules of ligand-protein domain associations were derived and successfully employed to predict protein targets outside the scope of the training set. This methodology was then tested on a proteomics affinity chromatography data set containing 699 compounds. The domain prediction model correctly detected 31.6% of the experimental targets at a specificity of 46.8%. This is striking since 86% of the predicted targets are not part of them (but share InterPro domains with them), and thus could not have been predicted by conventional target prediction approaches. Target predictions improve drastically when significance (FDR) scores for target pulldowns are employed, emphasizing their importance for eliminating artifacts. Filament proteins (such as actin and tubulin) are detected to be 'frequent hitters' in proteomics experiments and their presence in pulldowns is not supported by the target predictions. On the other hand, membrane-bound receptors such as serotonin and dopamine receptors are noticeably absent in the affinity chromatography sets, although their presence would be expected from the predicted targets of compounds. While this can partly be explained by the experimental setup, we suggest the computational methods employed here as a complementary step of identifying protein targets of small molecules. Affinity chromatography results for gefitinib are discussed in detail and while two out of the three kinases with the highest affinity to gefitinib in biochemical assays are detected by affinity chromatography, also the possible involvement of NSF as a target for modulating cancer progressions via beta-arrestin can be proposed by this method.

How similar are similarity searching methods? A principal component analysis of molecular descriptor space.

Different molecular descriptors capture different aspects of molecular structures, but this effect has not yet been quantified systematically on a large scale. In this work, we calculate the similarity of 37 descriptors by repeatedly selecting query compounds and ranking the rest of the database. Euclidean distances between the rank-ordering of different descriptors are calculated to determine descriptor (as opposed to compound) similarity, followed by PCA for visualization. Four broad descriptor classes are identified, which are circular fingerprints; circular fingerprints considering counts; path-based and keyed fingerprints; and pharmacophoric descriptors. Descriptor behavior is much more defined by those four classes than the particular parametrization. Using counts instead of the presence/absence of fingerprints significantly changes descriptor behavior, which is crucial for performance of topological autocorrelation vectors, but not circular fingerprints. Four-point pharmacophores (piDAPH4) surprisingly lead to much higher retrieval rates than three-point pharmacophores (28.21% vs 19.15%) but still similar rank-ordering of compounds (retrieval of similar actives). Looking into individual rankings, circular fingerprints seem more appropriate than path-based fingerprints if complex ring systems or branching patterns are present; count-based fingerprints could be more suitable in databases with a large number of repeated subunits (amide bonds, sugar rings, terpenes). Information-based selection of diverse fingerprints for consensus scoring (ECFP4/TGD fingerprints) led only to marginal improvement over single fingerprint results. While it seems to be nontrivial to exploit orthogonal descriptor behavior to improve retrieval rates in consensus virtual screening, those descriptors still each retrieve different actives which corroborates the strategy of employing diverse descriptors individually in prospective virtual screening settings.

Which aspects of HTS are empirically correlated with downstream success?

High-throughput screening (HTS) is a well-established hit-finding approach used in the pharmaceutical industry. In this article, recent experience at Novartis with respect to factors influencing the success of HTS campaigns is discussed. An inherent measure of HTS quality could be defined by the assay Z and Z' factors, the number of hits and their biological potencies; however, such measures of quality do not always correlate with the advancement of hits to the later stages of drug discovery. Also, for many target classes, such as kinases, it is easy to identify hits, but, as a result of selectivity, intellectual property and other issues, the projects do not result in lead declarations. In this article, HTS success is defined as the fraction of HTS campaigns that advance into the later stages of drug discovery, and the major influencing factors are examined. Interestingly, screening compounds in individual wells or in mixtures did not have a major impact on the HTS success and, equally interesting, there was no difference in the progression rates of biochemical and cell-based assays. Particular target types, assay technologies, structure-activity relationships and powder availability had a much greater impact on success as defined above. In addition, significant mutual dependencies can be observed - while one assay format works well with one target type, this situation might be completely reversed for a combination of the same readout technology with a different target type. The results and opinions presented here should be regarded as groundwork, and a plethora of factors that influence the fate of a project, such as biophysical measurements, chemical attractiveness of the hits, strategic reasons and safety pharmacology, are not covered here. Nonetheless, it is hoped that this information will be used industry-wide to improve success rates in terms of hits progressing into exploratory chemistry and beyond. The support that can be obtained from new in silico approaches to phase transitions are also described, along with the gaps they are designed to fill.

"Virtual fragment linking": an approach to identify potent binders from low affinity fragment hits.

In this work we explore the possibilities of using fragment-based screening data to prioritize compounds from a full HTS library, a method we call virtual fragment linking (VFL). The ability of VFL to identify compounds of nanomolar potency based on micromolar fragment binding data was tested on 75 target classes from the WOMBAT database and succeeded in 57 cases. Further, the method was demonstrated for seven drug targets from in-house screening programs that performed both FBS of 8800 fragments and screens of the full library. VFL captured between 28% and 67% of the hits (IC 50 < 10microM) in the top 5% of the ranked library for four of the targets (enrichment between 5-fold and 13-fold). Our findings lead us to conclude that proper coverage of chemical space by the fragment library is crucial for the VFL methodology to be successful in prioritizing HTS libraries from fragment-based screening data.

Entangled palladium nanoparticles in resin plugs.

Palladium nanoparticles were entrapped within resin plugs and used in a range of ligand-free cross-coupling reactions; the convenient modular format of the resin plug enhanced resin handling and allowed the catalysts to be easily recovered and multiply reused.

Chemogenomic data analysis: prediction of small-molecule targets and the advent of biological fingerprint.

Chemogenomics comprises a systematic relationship between targets and ligands that are used as target modulators in living systems such as cells or organisms. In recent years, data on small molecule-bioactivity relationships have become increasingly available, and consequently so have the number of approaches used to translate bioactivity data into knowledge. This review will focus on two aspects of chemogenomics. Firstly, in cases such as cell-based screens, the question of which target(s) a compound is modulating in order to cause the observed phenotype is crucial. In silico target prediction tools can suggest likely biological targets of small molecules via data mining in target-annotated chemical databases. This review presents some of the current tools available for this task and shows some sample applications relevant to a pharmaceutical industry setting. These applications are the prediction of false-positives in cell-based reporter gene assays, the prediction of targets by linking bioassay data with protein domain annotations, and the direct prediction of adverse reactions. Secondly, in recent years a shift from structure-derived chemical descriptors to biological descriptors has occurred. Here, the effect of a compound on a number of biological endpoints is used to make predictions about other properties, such as putative targets, associated adverse reactions, and pathways modulated by the compound. This review further summarizes these "performance" descriptors and their applications, focusing on gene expression profiles and high-content screening data. The advent of such biological fingerprints suggests that the field of drug discovery is currently at a crossroads, where single target bioassay results are supplanted by multidimensional biological fingerprints that reflect a new awareness of biological networks and polypharmacology.

Understanding false positives in reporter gene assays: in silico chemogenomics approaches to prioritize cell-based HTS data.

High throughput screening (HTS) data is often noisy, containing both false positives and negatives. Thus, careful triaging and prioritization of the primary hit list can save time and money by identifying potential false positives before incurring the expense of followup. Of particular concern are cell-based reporter gene assays (RGAs) where the number of hits may be prohibitively high to be scrutinized manually for weeding out erroneous data. Based on statistical models built from chemical structures of 650 000 compounds tested in RGAs, we created "frequent hitter" models that make it possible to prioritize potential false positives. Furthermore, we followed up the frequent hitter evaluation with chemical structure based in silico target predictions to hypothesize a mechanism for the observed "off target" response. It was observed that the predicted cellular targets for the frequent hitters were known to be associated with undesirable effects such as cytotoxicity. More specifically, the most frequently predicted targets relate to apoptosis and cell differentiation, including kinases, topoisomerases, and protein phosphatases. The mechanism-based frequent hitter hypothesis was tested using 160 additional druglike compounds predicted by the model to be nonspecific actives in RGAs. This validation was successful (showing a 50% hit rate compared to a normal hit rate as low as 2%), and it demonstrates the power of computational models toward understanding complex relations between chemical structure and biological function.

Analysis of pharmacology data and the prediction of adverse drug reactions and off-target effects from chemical structure.

Preclinical Safety Pharmacology (PSP) attempts to anticipate adverse drug reactions (ADRs) during early phases of drug discovery by testing compounds in simple, in vitro binding assays (that is, preclinical profiling). The selection of PSP targets is based largely on circumstantial evidence of their contribution to known clinical ADRs, inferred from findings in clinical trials, animal experiments, and molecular studies going back more than forty years. In this work we explore PSP chemical space and its relevance for the prediction of adverse drug reactions. Firstly, in silico (computational) Bayesian models for 70 PSP-related targets were built, which are able to detect 93% of the ligands binding at IC(50) < or = 10 microM at an overall correct classification rate of about 94%. Secondly, employing the World Drug Index (WDI), a model for adverse drug reactions was built directly based on normalized side-effect annotations in the WDI, which does not require any underlying functional knowledge. This is, to our knowledge, the first attempt to predict adverse drug reactions across hundreds of categories from chemical structure alone. On average 90% of the adverse drug reactions observed with known, clinically used compounds were detected, an overall correct classification rate of 92%. Drugs withdrawn from the market (Rapacuronium, Suprofen) were tested in the model and their predicted ADRs align well with known ADRs. The analysis was repeated for acetylsalicylic acid and Benperidol which are still on the market. Importantly, features of the models are interpretable and back-projectable to chemical structure, raising the possibility of rationally engineering out adverse effects. By combining PSP and ADR models new hypotheses linking targets and adverse effects can be proposed and examples for the opioid mu and the muscarinic M2 receptors, as well as for cyclooxygenase-1 are presented. It is hoped that the generation of predictive models for adverse drug reactions is able to help support early SAR to accelerate drug discovery and decrease late stage attrition in drug discovery projects. In addition, models such as the ones presented here can be used for compound profiling in all development stages.