Assessing machine learning approaches for predicting failures of investigational drug candidates during clinical trials.

One of the major challenges in drug development is having acceptable levels of efficacy and safety throughout all the phases of clinical trials followed by the successful launch in the market. While there are many factors such as molecular properties, toxicity parameters, mechanism of action at the target site, etc. that regulates the therapeutic action of a compound, a holistic approach directed towards data-driven studies will invariably strengthen the predictive toxicological sciences. Our quest for the current study is to find out various reasons as to why an investigational candidate would fail in the clinical trials after multiple iterations of refinement and optimization. We have compiled a dataset that comprises of approved and withdrawn drugs as well as toxic compounds and essentially have used time-split based approach to generate the training and validation set. Five highly robust and scalable machine learning binary classifiers were used to develop the predictive models that were trained with features like molecular descriptors and fingerprints and then validated rigorously to achieve acceptable performance in terms of a set of performance metrics. The mean AUC scores for all the five classifiers with the hold-out test set were obtained in the range of 0.66-0.71. The models were further used to predict the probability score for the clinical candidate dataset. The top compounds predicted to be toxic were analyzed to estimate different dimensions of toxicity. Apparently, through this study, we propose that with the appropriate use of feature extraction and machine learning methods, one can estimate the likelihood of success or failure of investigational drugs candidates thereby opening an avenue for future trends in computational toxicological studies. The models developed in the study can be accessed at https://github.com/gnsastry/predicting_clinical_trials.git.

Bis(porphyrin)-anthraquinone triads: synthesis, spectroscopy, and photochemistry.

Molecular triads based on bis(porphyrin)-anthraquinone having azomethine bridge at the pyrrole-ß position have been designed and synthesized. Both free-base AQ-(H2)2 and zinc AQ-(Zn)2 triads are characterized by elemental analysis, MALDI-MS, (1)H NMR, UV-visible, and fluorescence spectroscopy (steady-state and time-resolved) as well as electrochemical method. The absorption spectra of both Soret and Q-bands of the triads are red-shifted by 12-20 nm with respect to their monomer units. The study supported by theoretical calculations manifests that there exists a negligible electronic communication in the ground state between the donor porphyrin and acceptor anthraquinone of these triads. However, interestingly, both the triads exhibit significant fluorescence emission quenching (51-92%) of the porphyrin emission compared to their monomeric units. The emission quenching is attributed to the excited-state intramolecular photoinduced electron transfer from porphyrins to anthraquinone. The electron-transfer rates (kET) of these triads are found in the range 1.0 × 10(8) to 7.7 × 10(9) s(-1) and are found to be solvent dependent.

Differentiation of diastereomeric N-aryltetrahydropyrano/tetrahydrofuranochromenylamines under electron ionization and chemical ionization conditions.

A series of diastereomeric 4S,5S,6R/S-tetrahydropyrano- and 3S,4S,5R/S-tetrahydrofuranochromenylamine derivatives (a/b isomers; 1-26) has been studied under electron ionization (EI) and chemical ionization (CI) conditions. The EI mass spectra of all diastereomeric compounds show two characteristic fragment ions, of which one is formed by retro-Diels-Alder (RDA) reaction from the molecular ion, retaining the charge on the diene fragment, and the other [M-(HNAr)]+ ion by a simple radical loss. The RDA process is more favorable in all b isomers, whereas the radical loss is dominant in all a isomers; based on these two ions it is easy to differentiate the two diastereomers. The collision-induced dissociation (CID) spectra of all the molecular ions also show the same trend, which reflects the stereoselectivity in the formation of the two characteristic fragment ions. The results of theoretical calculations performed are in accordance with the experimental observations. The CI experiments (methane and isobutane) on all the diastereomeric compounds also enabled the differentiation of the isomers.

Differentiation of diastereomeric conduramine derivatives under electron ionization and chemical ionization mass spectral conditions.

Diastereomeric conduramine derivatives, i.e., (1R,2S,3R/S,6S)-6-(N-carbomethoxyamino) 1,2-O-isopropylidenecyclohex-4-ene-1,2,3-triol (1 and 2) and their O-acetyl derivatives (3 and 4), were studied using gas chromatography (GC) with electron ionization (EI) and chemical ionization (CI). The EI mass spectra of diastereomeric pairs show consistent differences in the relative abundances of characteristic ions. The EI fragmentation patterns are based on precursor/product ion spectra, high-resolution mass spectrometry (HRMS) and deuterium labelling. The CI spectra show differences from the EI spectra, and the isobutane/CI spectra are much simpler than the methane/CI spectra. The differences shown in the CI spectra are similar to those shown in the product ion spectra of [M+H](+) ions generated under electrospray ionization (ESI) conditions. Theoretical calculations are performed to understand the observed differences. The differences in the relative stabilities of molecular ions, or protonated molecules at different sites, can explain the observed differences in the spectra.