Relation extraction from DailyMed structured product labels by optimally combining crowd, experts and machines.

The effectiveness of machine learning models to provide accurate and consistent results in drug discovery and clinical decision support is strongly dependent on the quality of the data used. However, substantive amounts of open data that drive drug discovery suffer from a number of issues including inconsistent representation, inaccurate reporting, and incomplete context. For example, databases of FDA-approved drug indications used in computational drug repositioning studies do not distinguish between treatments that simply offer symptomatic relief from those that target the underlying pathology. Moreover, drug indication sources often lack proper provenance and have little overlap. Consequently, new predictions can be of poor quality as they offer little in the way of new insights. Hence, work remains to be done to establish higher quality databases of drug indications that are suitable for use in drug discovery and repositioning studies. Here, we report on the combination of weak supervision (i.e., programmatic labeling and crowdsourcing) and deep learning methods for relation extraction from DailyMed text to create a higher quality drug-disease relation dataset. The generated drug-disease relation data shows a high overlap with DrugCentral, a manually curated dataset. Using this dataset, we constructed a machine learning model to classify relations between drugs and diseases from text into four categories; treatment, symptomatic relief, contradiction, and effect, exhibiting an improvement of 15.5% with Bi-LSTM (F1 score of 71.8%) over the best performing discrete method. Access to high quality data is crucial to building accurate and reliable drug repurposing prediction models. Our work suggests how the combination of crowds, experts, and machine learning methods can go hand-in-hand to improve datasets and predictive models.

Enantioselectivity Induced by Stereoselective Interlocking: A Novel Core Motif for Tropos Ligands.

Well-defined supramolecular interactions are a powerful tool to control the stereochemistry of a catalytic reaction. In this paper, we report a novel core motif for fluxional 2,2'-biphenyl ligands carrying (S)-amino acid-derived interaction sites in 5,5'-position that cause spontaneous enrichment of the Rax rotamer. The process is based on strong non-covalent interlocking between interaction sites, which causes diastereoselective formation of a supramolecular ligand dimer, in which the axial chirality of the two subunits is dictated by the stereochemical information in the amino acid residues. The detailed structure of the dimer was elucidated by NMR spectroscopy and single-crystal X-ray analysis. Three different phosphorus-based ligand types, namely a bisphosphine, a bisphosphinite and a phosphoramidite were synthesized and characterized. Whereas the first one was found to exist in a strongly weighted equilibrium, the two others each exhibited stereoconvergent behavior transforming into the diastereopure Rax rotamer. Enriched ligands were used in rhodium-mediated asymmetric hydrogenation reactions of prochiral olefins in which very high enantioselectivities of up to 96:4 were achieved.

Design and synthesis of a stereodynamic catalyst with reversal of selectivity by enantioselective self-inhibition.

Chirality plays a pivotal role in an uncountable number of biological processes, and nature has developed intriguing mechanisms to maintain this state of enantiopurity. The strive for a deeper understanding of the different elements that constitute such self-sustaining systems on a molecular level has sparked great interest in the studies of autoinductive and amplifying enantioselective reactions. The design of these reactions remains highly challenging; however, the development of generally applicable principles promises to have a considerable impact on research of catalyst design and other adjacent fields in the future. Here, we report the realization of an autoinductive, enantioselective self-inhibiting hydrogenation reaction. Development of a stereodynamic catalyst with chiral sensing abilities allowed for a chiral reaction product to interact with the catalyst and change its selectivity in order to suppress its formation, which caused a reversal of selectivity over time.

Inducing Enantioselectivity in a Dynamic Catalyst by Supramolecular Interlocking.

The design of a new class of fluxional biphenyl bisphosphinite (BIBIPHOS) ligands decorated with amino acid-based diamide interaction sites is reported that undergo spontaneous desymmetrization. Hydrogenation of prochiral alkenes using Rh-BIBIPHOS results in enantiomeric ratios of up to 96:4 (R/S). This stereoconvergent behavior of the fluxional BIBIPHOS ligand is triggered by pronounced intermolecular interlocking of the recognition sites, leading to the formation of a supramolecular assembly, where the axial orientation of the biphenyl ligand backbone is governed by the chirality of the amino acid moieties. Stereoinduction during catalysis is decoupled from this process and occurs as an immediate consequence of the emergent behavior of the ligands. This supramolecular system is very robust and has the potential to be adopted for other ligand designs in enantioselective catalysis.

Trends in covalent drug discovery: a 2020-23 patent landscape analysis focused on select covalent reacting groups (CRGs) found in FDA-approved drugs.

INTRODUCTION: Covalent drugs contain electrophilic groups that can react with nucleophilic amino acids located in the active sites of proteins, particularly enzymes. Recently, there has been considerable interest in using covalent drugs to target non-catalytic amino acids in proteins to modulate difficult targets (i.e. targeted covalent inhibitors). Covalent compounds contain a wide variety of covalent reacting groups (CRGs), but only a few of these CRGs are present in FDA-approved covalent drugs. AREAS COVERED: This review summarizes a 2020-23 patent landscape analysis that examined trends in the field of covalent drug discovery around targets and organizations. The analysis focused on patent applications that were submitted to the World International Patent Organization and selected using a combination of keywords and structural searches based on CRGs present in FDA-approved drugs. EXPERT OPINION: A total of 707 patent applications from >300 organizations were identified, disclosing compounds that acted at 71 targets. Patent application counts for five targets accounted for ~63% of the total counts (i.e. BTK, EGFR, FGFR, KRAS, and SARS-CoV-2 Mpro). The organization with the largest number of patent counts was an academic institution (Dana-Farber Cancer Institute). For one target, KRAS G12C, the discovery of new drugs was highly competitive (>100 organizations, 186 patent applications).