Migrating to Long-Read Sequencing for Clinical Routine BCR-ABL1 TKI Resistance Mutation Screening.

Objective: The aim of this project was to implement long-read sequencing for BCR-ABL1 TKI resistance mutation screening in a clinical setting for patients undergoing treatment for chronic myeloid leukemia. Materials and Methods: Processes were established for registering and transferring samples from the clinic to an academic sequencing facility for long-read sequencing. An automated analysis pipeline for detecting mutations was established, and an information system was implemented comprising features for data management, analysis and visualization. Clinical validation was performed by identifying BCR-ABL1 TKI resistance mutations by Sanger and long-read sequencing in parallel. The developed software is available as open source via GitHub at https://github.com/pharmbio/clamp. Results: The information system enabled traceable transfer of samples from the clinic to the sequencing facility, robust and automated analysis of the long-read sequence data, and communication of results from sequence analysis in a reporting format that could be easily interpreted and acted upon by clinical experts. In a validation study, all 17 resistance mutations found by Sanger sequencing were also detected by long-read sequencing. An additional 16 mutations were found only by long-read sequencing, all of them with frequencies below the limit of detection for Sanger sequencing. The clonal distributions of co-existing mutations were automatically resolved through the long-read data analysis. After the implementation and validation, the clinical laboratory switched their routine protocol from using Sanger to long-read sequencing for this application. Conclusions: Long-read sequencing delivers results with higher sensitivity compared to Sanger sequencing and enables earlier detection of emerging TKI resistance mutations. The developed processes, analysis workflow, and software components lower barriers for adoption and could be extended to other applications.

Predicting protein network topology clusters from chemical structure using deep learning.

Comparing chemical structures to infer protein targets and functions is a common approach, but basing comparisons on chemical similarity alone can be misleading. Here we present a methodology for predicting target protein clusters using deep neural networks. The model is trained on clusters of compounds based on similarities calculated from combined compound-protein and protein-protein interaction data using a network topology approach. We compare several deep learning architectures including both convolutional and recurrent neural networks. The best performing method, the recurrent neural network architecture MolPMoFiT, achieved an F1 score approaching 0.9 on a held-out test set of 8907 compounds. In addition, in-depth analysis on a set of eleven well-studied chemical compounds with known functions showed that predictions were justifiable for all but one of the chemicals. Four of the compounds, similar in their molecular structure but with dissimilarities in their function, revealed advantages of our method compared to using chemical similarity.

Rapid Increase in Carriage Rates of Enterobacteriaceae Producing Extended-Spectrum ß-Lactamases in Healthy Preschool Children, Sweden.

By collecting and analyzing diapers, we identified a >6-fold increase in carriage of extended-spectrum ß-lactamase (ESBL)-producing Enterobacteriaceae for healthy preschool children in Sweden (p<0.0001). For 6 of the 50 participating preschools, the carriage rate was >40%. We analyzed samples from 334 children and found 56 containing >1 ESBL producer. The prevalence in the study population increased from 2.6% in 2010 to 16.8% in 2016 (p<0.0001), and for 6 of the 50 participating preschools, the carriage rate was >40%. Furthermore, 58% of the ESBL producers were multidrug resistant, and transmission of ESBL-producing and non-ESBL-producing strains was observed at several of the preschools. Toddlers appear to be major carriers of ESBL producers in Sweden.

A confidence predictor for logD using conformal regression and a support-vector machine.

Lipophilicity is a major determinant of ADMET properties and overall suitability of drug candidates. We have developed large-scale models to predict water-octanol distribution coefficient (logD) for chemical compounds, aiding drug discovery projects. Using ACD/logD data for 1.6 million compounds from the ChEMBL database, models are created and evaluated by a support-vector machine with a linear kernel using conformal prediction methodology, outputting prediction intervals at a specified confidence level. The resulting model shows a predictive ability of [Formula: see text] and with the best performing nonconformity measure having median prediction interval of [Formula: see text] log units at 80% confidence and [Formula: see text] log units at 90% confidence. The model is available as an online service via an OpenAPI interface, a web page with a molecular editor, and we also publish predictive values at 90% confidence level for 91 M PubChem structures in RDF format for download and as an URI resolver service.

Tracking the NGS revolution: managing life science research on shared high-performance computing clusters.

Background: Next-generation sequencing (NGS) has transformed the life sciences, and many research groups are newly dependent upon computer clusters to store and analyze large datasets. This creates challenges for e-infrastructures accustomed to hosting computationally mature research in other sciences. Using data gathered from our own clusters at UPPMAX computing center at Uppsala University, Sweden, where core hour usage of ∼800 NGS and ∼200 non-NGS projects is now similar, we compare and contrast the growth, administrative burden, and cluster usage of NGS projects with projects from other sciences. Results: The number of NGS projects has grown rapidly since 2010, with growth driven by entry of new research groups. Storage used by NGS projects has grown more rapidly since 2013 and is now limited by disk capacity. NGS users submit nearly twice as many support tickets per user, and 11 more tools are installed each month for NGS projects than for non-NGS projects. We developed usage and efficiency metrics and show that computing jobs for NGS projects use more RAM than non-NGS projects, are more variable in core usage, and rarely span multiple nodes. NGS jobs use booked resources less efficiently for a variety of reasons. Active monitoring can improve this somewhat. Conclusions: Hosting NGS projects imposes a large administrative burden at UPPMAX due to large numbers of inexperienced users and diverse and rapidly evolving research areas. We provide a set of recommendations for e-infrastructures that host NGS research projects. We provide anonymized versions of our storage, job, and efficiency databases.

Efficient iterative virtual screening with Apache Spark and conformal prediction.

BACKGROUND: Docking and scoring large libraries of ligands against target proteins forms the basis of structure-based virtual screening. The problem is trivially parallelizable, and calculations are generally carried out on computer clusters or on large workstations in a brute force manner, by docking and scoring all available ligands. CONTRIBUTION: In this study we propose a strategy that is based on iteratively docking a set of ligands to form a training set, training a ligand-based model on this set, and predicting the remainder of the ligands to exclude those predicted as 'low-scoring' ligands. Then, another set of ligands are docked, the model is retrained and the process is repeated until a certain model efficiency level is reached. Thereafter, the remaining ligands are docked or excluded based on this model. We use SVM and conformal prediction to deliver valid prediction intervals for ranking the predicted ligands, and Apache Spark to parallelize both the docking and the modeling. RESULTS: We show on 4 different targets that conformal prediction based virtual screening (CPVS) is able to reduce the number of docked molecules by 62.61% while retaining an accuracy for the top 30 hits of 94% on average and a speedup of 3.7. The implementation is available as open source via GitHub ( https://github.com/laeeq80/spark-cpvs ) and can be run on high-performance computers as well as on cloud resources.

Large-scale virtual screening on public cloud resources with Apache Spark.

BACKGROUND: Structure-based virtual screening is an in-silico method to screen a target receptor against a virtual molecular library. Applying docking-based screening to large molecular libraries can be computationally expensive, however it constitutes a trivially parallelizable task. Most of the available parallel implementations are based on message passing interface, relying on low failure rate hardware and fast network connection. Google's MapReduce revolutionized large-scale analysis, enabling the processing of massive datasets on commodity hardware and cloud resources, providing transparent scalability and fault tolerance at the software level. Open source implementations of MapReduce include Apache Hadoop and the more recent Apache Spark. RESULTS: We developed a method to run existing docking-based screening software on distributed cloud resources, utilizing the MapReduce approach. We benchmarked our method, which is implemented in Apache Spark, docking a publicly available target receptor against [Formula: see text]2.2 M compounds. The performance experiments show a good parallel efficiency (87%) when running in a public cloud environment. CONCLUSION: Our method enables parallel Structure-based virtual screening on public cloud resources or commodity computer clusters. The degree of scalability that we achieve allows for trying out our method on relatively small libraries first and then to scale to larger libraries. Our implementation is named Spark-VS and it is freely available as open source from GitHub (https://github.com/mcapuccini/spark-vs).Graphical abstract.

Docking of Macrocycles: Comparing Rigid and Flexible Docking in Glide.

In recent years, there has been an increased interest in using macrocyclic compounds for drug discovery and development. For docking of these commonly large and flexible compounds to be addressed, a screening and a validation set were assembled from the PDB consisting of 16 and 31 macrocycle-containing protein complexes, respectively. The macrocycles were docked in Glide by rigid docking of pregenerated conformational ensembles produced by the macrocycle conformational sampling method (MCS) in Schrödinger Release 2015-3 or by direct Glide flexible docking after performing ring-templating. The two protocols were compared to rigid docking of pregenerated conformational ensembles produced by an exhaustive Monte Carlo multiple minimum (MCMM) conformational search and a shorter MCMM conformational search (MCMM-short). The docking accuracy was evaluated and expressed as the RMSD between the heavy atoms of the ligand as found in the X-ray structure after refinement and the poses obtained by the docking protocols. The median RMSD values for top-scored poses of the screening set were 0.83, 0.80, 0.88, and 0.58 Å for MCMM, MCMM-short, MCS, and Glide flexible docking, respectively. There was no statistically significant difference in the performance between rigid docking of pregenerated conformations produced by the MCS and direct docking using Glide flexible docking. However, the flexible docking protocol was 2-times faster in docking the screening set compared to that of the MCS protocol. In a final study, the new Prime-MCS method was evaluated in Schrödinger Release 2016-3. This method is faster compared that of to MCS; however, the conformations generated were found to be suboptimal for rigid docking. Therefore, on the basis of timing, accuracy, and ease of set up, standard Glide flexible docking with prior ring-templating is recommended over current gold standard protocols using rigid docking of pregenerated conformational ensembles.

Large-scale ligand-based predictive modelling using support vector machines.

The increasing size of datasets in drug discovery makes it challenging to build robust and accurate predictive models within a reasonable amount of time. In order to investigate the effect of dataset sizes on predictive performance and modelling time, ligand-based regression models were trained on open datasets of varying sizes of up to 1.2 million chemical structures. For modelling, two implementations of support vector machines (SVM) were used. Chemical structures were described by the signatures molecular descriptor. Results showed that for the larger datasets, the LIBLINEAR SVM implementation performed on par with the well-established libsvm with a radial basis function kernel, but with dramatically less time for model building even on modest computer resources. Using a non-linear kernel proved to be infeasible for large data sizes, even with substantial computational resources on a computer cluster. To deploy the resulting models, we extended the Bioclipse decision support framework to support models from LIBLINEAR and made our models of logD and solubility available from within Bioclipse.

N-Substituted pyrazole-3-carboxamides as inhibitors of human 15-lipoxygenase.

High-throughput screening was used to find selective inhibitors of human 15-lipoxygenase-1 (15-LOX-1). One hit, a 1-benzoyl substituted pyrazole-3-carboxanilide (1a), was used as a starting point in a program to develop potent and selective 15-LOX-1 inhibitors.

3-Substituted pyrazoles and 4-substituted triazoles as inhibitors of human 15-lipoxygenase-1.

Investigation of 1N-substituted pyrazole-3-carboxanilides as 15-lipoxygenase-1 (15-LOX-1) inhibitors demonstrated that the 1N-substituent was not essential for activity or selectivity. Additional halogen substituents on the pyrazole ring, however, increased activity. Further development led to triazole-4-carboxanilides and 2-(3-pyrazolyl) benzoxazoles, which are potent and selective 15-LOX-1 inhibitors.

Toward a benchmarking data set able to evaluate ligand- and structure-based virtual screening using public HTS data.

Virtual screening has the potential to accelerate and reduce costs of probe development and drug discovery. To develop and benchmark virtual screening methods, validation data sets are commonly used. Over the years, such data sets have been constructed to overcome the problems of analogue bias and artificial enrichment. With the rapid growth of public domain databases containing high-throughput screening data, such as the PubChem BioAssay database, there is an increased possibility to use such data for validation. In this study, we identify PubChem data sets suitable for validation of both structure- and ligand-based virtual screening methods. To achieve this, high-throughput screening data for which a crystal structure of the bioassay target was available in the PDB were identified. Thereafter, the data sets were inspected to identify structures and data suitable for use in validation studies. In this work, we present seven data sets (MMP13, DUSP3, PTPN22, EPHX2, CTDSP1, MAPK10, and CDK5) compiled using this method. In the seven data sets, the number of active compounds varies between 19 and 369 and the number of inactive compounds between 59 405 and 337 634. This gives a higher ratio of the number of inactive to active compounds than what is found in most benchmark data sets. We have also evaluated the screening performance using docking and 3D shape similarity with default settings. To characterize the data sets, we used physicochemical similarity and 2D fingerprint searches. We envision that these data sets can be a useful complement to current data sets used for method evaluation.

Automated QuantMap for rapid quantitative molecular network topology analysis.

SUMMARY: The previously disclosed QuantMap method for grouping chemicals by biological activity used online services for much of the data gathering and some of the numerical analysis. The present work attempts to streamline this process by using local copies of the databases and in-house analysis. Using computational methods similar or identical to those used in the previous work, a qualitatively equivalent result was found in just a few seconds on the same dataset (collection of 18 drugs). We use the user-friendly Galaxy framework to enable users to analyze their own datasets. Hopefully, this will make the QuantMap method more practical and accessible and help achieve its goals to provide substantial assistance to drug repositioning, pharmacology evaluation and toxicology risk assessment. AVAILABILITY: http://galaxy.predpharmtox.org CONTACT: mats.gustafsson@medsci.uu.se or ola.spjuth@farmbio.uu.se SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Development of a fluorescent intensity assay amenable for high-throughput screening for determining 15-lipoxygenase activity.

15-Lipoxygenase-1 catalyzes the introduction of molecular oxygen into polyunsaturated fatty acids to form a lipid hydroperoxide. The authors have developed an assay for the detection of lipid hydroperoxides formed by human 15-lipoxygenase (15-LO) in enzyme or cellular assays using either a 96-well or a 384-well format. The assays described take advantage of the ability of lipid hydroperoxides to oxidize nonfluorescent diphenyl-1-pyrenylphosphine (DPPP) to a fluorescent phosphine oxide. Oxidation of DPPP yields a fluorescent compound, which is not sensitive to temperature and is stable for more than 2 h. The assay is sensitive toward inhibition and robust with a Z' value of 0.79 and 0.4 in a 96- and 384-well format, respectively, and thus amenable for high-throughput screening. The utility of DPPP as a marker for 15-lipoxygenase activity was demonstrated with both enzyme- and cell-based assays for the identification of hits and to determine potency by IC(50) determinations.

Improved CoMFA modeling by optimization of settings.

The possibility of improving the predictive ability of comparative molecular field analysis (CoMFA) by settings optimization has been evaluated to show that CoMFA predictive ability can be improved. Ten different CoMFA settings are evaluated, producing a total of 6120 models. This method has been applied to nine different data sets, including the widely used benchmark steroid data set, as well as eight other data sets proposed as QSAR benchmarking data sets by Sutherland et al. (J. Med. Chem. 2004, 47, 5541-5554). All data sets have been studied using training and test sets to allow for both internal (q(2)) and external (r(2)(pred)) predictive ability assessment. CoMFA settings optimization was successful in developing models with improved q(2) and r(2)(pred) as compared to default CoMFA modeling. Optimized CoMFA is compared with comparative molecular similarity indices analysis (CoMSIA) and holographic quantitative structure-activity relationship (HQSAR) models and found to consistently produce models with improved or equivalent q(2) and r(2)(pred). The ability of settings optimization to improve model predictive ability has been validated using both internal and external predictions, and the risk of chance correlation has been evaluated using response variable randomization tests.

Cyclic sulfamide HIV-1 protease inhibitors, with sidechains spanning from P2/P2' to P1/P1'.

Previous studies of HIV protease inhibitors have shown that it is possible to elongate the P1/P1' sidechains to reach the S3/S3' binding sites. By analogy, we expected that it would be possible to design inhibitors reaching between the S1/S1' and S2/S2' binding sites. Molecular modeling suggested that this could be achieved with appropriate ortho-substitution of the P2/P2' benzyl groups in our cyclic sulfamide inhibitors. Four different spacer groups were investigated. The compounds were smoothly prepared from tartaric acid in five steps and exhibit low to moderate activity, the most potent inhibitor possessing a Ki value of 0.53 microM.

High-speed optimization of inhibitors of the malarial proteases plasmepsin I and II.

Four focused libraries targeted for inhibition of the malarial proteases plasmepsin I and II were designed, synthesized, purified, and screened. Selected carboxylic acids and organometallic reactants with diverse physical properties were attached to the hydroxylethylamine scaffold in the P3 and P1' positions to furnish inhibitors with highly improved activity. The concept of controlled and sequential microwave heating was employed for rapid library generation. This combinatorial optimization protocol afforded plasmepsin inhibitors not only with K(i) values in the low nanomolar range, but also with high selectivity versus the human protease cathepsin D. With this class of inhibitory agents, modifications of the P1' substituents resulted in the largest impact on the plasmepsin/cathepsin D selectivity.

Optimization of P1-P3 groups in symmetric and asymmetric HIV-1 protease inhibitors.

HIV-1 protease is an important target for treatment of AIDS, and efficient drugs have been developed. However, the resistance and negative side effects of the current drugs has necessitated the development of new compounds with different binding patterns. In this study, nine C-terminally duplicated HIV-1 protease inhibitors were cocrystallised with the enzyme, the crystal structures analysed at 1.8-2.3 A resolution, and the inhibitory activity of the compounds characterized in order to evaluate the effects of the individual modifications. These compounds comprise two central hydroxy groups that mimic the geminal hydroxy groups of a cleavage-reaction intermediate. One of the hydroxy groups is located between the delta-oxygen atoms of the two catalytic aspartic acid residues, and the other in the gauche position relative to the first. The asymmetric binding of the two central inhibitory hydroxyls induced a small deviation from exact C2 symmetry in the whole enzyme-inhibitor complex. The study shows that the protease molecule could accommodate its structure to different sizes of the P2/P2' groups. The structural alterations were, however, relatively conservative and limited. The binding capacity of the S3/S3' sites was exploited by elongation of the compounds with groups in the P3/P3' positions or by extension of the P1/P1' groups. Furthermore, water molecules were shown to be important binding links between the protease and the inhibitors. This study produced a number of inhibitors with Ki values in the 100 picomolar range.

Relationships between structure and interaction kinetics for HIV-1 protease inhibitors.

The interaction between HIV-1 protease and 58 structurally diverse transition-state analogue inhibitors has been analyzed by a surface plasmon resonance based biosensor. Association and dissociation rate constants and affinities were determined and displayed as k(on)-k(off)-K(D) maps. It was shown that different classes of inhibitors fall into distinct clusters in these maps. Significant changes in association and dissociation rates were found as a result of modifying the P1/P1' or P2/P2' side chains of a linear lead compound. Similarly, cyclic urea and cyclic sulfamide inhibitors displayed different kinetic features and the affinities of both classes of cyclic compounds were limited by fast dissociation rates. These results confirm that association and dissociation rates are important features of drug-target interactions and indicate that optimization of inhibitor efficacy may be guided by aiming for high association and low dissociation rates rather than high affinity alone. The present approach thus provides a new tool for structure-interaction kinetic analysis and drug discovery.