FEP Augmentation as a Means to Solve Data Paucity Problems for Machine Learning in Chemical Biology.

In the realm of medicinal chemistry, the primary objective is to swiftly optimize a multitude of chemical properties of a set of compounds to yield a clinical candidate poised for clinical trials. In recent years, two computational techniques, machine learning (ML) and physics-based methods, have evolved substantially and are now frequently incorporated into the medicinal chemist's toolbox to enhance the efficiency of both hit optimization and candidate design. Both computational methods come with their own set of limitations, and they are often used independently of each other. ML's capability to screen extensive compound libraries expediently is tempered by its reliance on quality data, which can be scarce especially during early-stage optimization. Contrarily, physics-based approaches like free energy perturbation (FEP) are frequently constrained by low throughput and high cost by comparison; however, physics-based methods are capable of making highly accurate binding affinity predictions. In this study, we harnessed the strength of FEP to overcome data paucity in ML by generating virtual activity data sets which then inform the training of algorithms. Here, we show that ML algorithms trained with an FEP-augmented data set could achieve comparable predictive accuracy to data sets trained on experimental data from biological assays. Throughout the paper, we emphasize key mechanistic considerations that must be taken into account when aiming to augment data sets and lay the groundwork for successful implementation. Ultimately, the study advocates for the synergy of physics-based methods and ML to expedite the lead optimization process. We believe that the physics-based augmentation of ML will significantly benefit drug discovery, as these techniques continue to evolve.

What is missing in our understanding of urban slum environments and maternal, infant and young child nutrition from publicly available data in Asia and the Pacific?

Given the recent, rapid urbanisation in Asia and the Pacific region, coupled with increases in the triple burden of malnutrition, we need to better understand maternal, infant and young child nutrition (MIYCN) for populations living in urban slum environments. This research used existing large-scale datasets to explore MIYCN indicators for those living in urban slum, compared with urban nonslum, areas. Data since 2015 from available Demographic and Health Surveys (DHS; Afghanistan, India, Indonesia, Myanmar, Pakistan and the Philippines) and Multiple Indicator Cluster Surveys (MICS; Bangladesh, Fiji, Kiribati, Mongolia, Nepal, Thailand and Tuvalu) were analysed. Most urban children in the 13 countries from the region were breastfed within 24 h of birth, with slightly higher rates for those living in slums. Conversely, almost all indicators of infant and young child malnutrition were worse for those in urban slums. For mothers living in slums, underweight prevalence and iron deficiency anaemia were higher while maternal overweight and obesity prevalence were lower. Analysis revealed disparities across countries in the wealth status of those living in slum versus nonslum areas. What is currently missing is representative sampling of households, adequate collection of data both within and across countries, and accurate representation of slum-dwellers in large-scale surveys. Given that limited data for the region show urban poor slum populations are vulnerable to poor nutrition indicators, more data are needed on the poorest urban slum populations to enable effective resource allocation to support optimal MIYCN.

Fake it until you make it? Generative de novo design and virtual screening of synthesizable molecules.

Computational techniques, including virtual screening, de novo design, and generative models, play an increasing role in expediting DMTA cycles for modern molecular discovery. However, computationally proposed molecules must be synthetically feasible for laboratory testing. In this perspective, we offer a succinct introduction to the subject, and showcase typical workflows to integrate synthesis planning, synthesizability scoring, and molecule generation. Finally, we address limitations and opportunities for future research.

PREFER: A New Predictive Modeling Framework for Molecular Discovery.

Machine-learning and deep-learning models have been extensively used in cheminformatics to predict molecular properties, to reduce the need for direct measurements, and to accelerate compound prioritization. However, different setups and frameworks and the large number of molecular representations make it difficult to properly evaluate, reproduce, and compare them. Here we present a new PREdictive modeling FramEwoRk for molecular discovery (PREFER), written in Python (version 3.7.7) and based on AutoSklearn (version 0.14.7), that allows comparison between different molecular representations and common machine-learning models. We provide an overview of the design of our framework and show exemplary use cases and results of several representation-model combinations on diverse data sets, both public and in-house. Finally, we discuss the use of PREFER on small data sets. The code of the framework is freely available on GitHub.

Diet and food insecurity among mothers, infants, and young children in Peru before and during COVID-19: A panel survey.

The COVID-19 pandemic may impact diet and nutrition through increased household food insecurity, lack of access to health services, and poorer quality diets. The primary aim of this study is to assess the impact of the pandemic on dietary outcomes of mothers and their infants and young children (IYC) in low-income urban areas of Peru. We conducted a panel study, with one survey prepandemic (n = 244) and one survey 9 months after the onset of COVID-19 (n = 254). We assessed breastfeeding and complementary feeding indicators and maternal dietary diversity in both surveys. During COVID-19, we assessed household food insecurity experience and economic impacts of the pandemic on livelihoods; receipt of financial or food assistance, and uptake of health services. Almost all respondents (98.0%) reported adverse economic impacts due to the pandemic and 46.9% of households were at risk of moderate or severe household food insecurity. The proportion of households receiving government food assistance nearly doubled between the two surveys (36.5%-59.5%). Dietary indicators, however, did not worsen in mothers or IYC. Positive changes included an increase in exclusive breastfeeding <6 months (24.2%-39.0%, p < 0.008) and a decrease in sweet food consumption by IYC (33.1%-18.1%, p = 0.001) and mothers (34.0%-14.6%, p < 0.001). The prevalence of sugar-sweetened beverage consumption remained high in both mothers (97%) and IYC (78%). In sum, we found dietary indicators had not significantly worsened 9 months into the COVID-19 pandemic. However, several indicators remain suboptimal and should be targeted in future interventions.

An investigation of coping in response to different race-related stressor experiences in school among racially diverse participants.

The landmark Brown versus Board of Education decision led to the desegregation of public schools in the United States. Consequently, Black/African American, Latinx, Asian, and Native American (BALANA) students experienced multiple race-related stressors. Not surprisingly, BALANA students still report experiences with racial discrimination, exclusion, and harassment in public schools. Encounters with race-related stressors in school can lead to maladaptive outcomes; however, for some young people, such stressors may induce a more adaptive coping response. The coping response may depend on the experience in school and the degree to which young people perceive they have control over the stressor. This study focused on the transactional nature of coping concerning different race-related stressors in school. The study investigates whether emotion-focused coping use prevailed across a sample of 165 ethnically and racially diverse participants. Results suggest that emotion-focused coping use is high among participants; however, the unique race-related stressor experienced in school can also lead to social-focused and other coping strategies. Coping with race-related stressors in school allows BALANA students to survive racism; reducing the presence of race-related stressors in school allows BALANA students to thrive. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Single-photon emission from single-electron transport in a SAW-driven lateral light-emitting diode.

The long-distance quantum transfer between electron-spin qubits in semiconductors is important for realising large-scale quantum computing circuits. Electron-spin to photon-polarisation conversion is a promising technology for achieving free-space or fibre-coupled quantum transfer. In this work, using only regular lithography techniques on a conventional 15 nm GaAs quantum well, we demonstrate acoustically-driven generation of single photons from single electrons, without the need for a self-assembled quantum dot. In this device, a single electron is carried in a potential minimum of a surface acoustic wave (SAW) and is transported to a region of holes to form an exciton. The exciton then decays and creates a single optical photon within 100 ps. This SAW-driven electroluminescence, without optimisation, yields photon antibunching with g(2)(0) = 0.39 ± 0.05 in the single-electron limit (g(2)(0) = 0.63 ± 0.03 in the raw histogram). Our work marks the first step towards electron-to-photon (spin-to-polarisation) qubit conversion for scaleable quantum computing architectures.

Controlling the coherence of a diamond spin qubit through its strain environment.

The uncontrolled interaction of a quantum system with its environment is detrimental for quantum coherence. For quantum bits in the solid state, decoherence from thermal vibrations of the surrounding lattice can typically only be suppressed by lowering the temperature of operation. Here, we use a nano-electro-mechanical system to mitigate the effect of thermal phonons on a spin qubit - the silicon-vacancy colour centre in diamond - without changing the system temperature. By controlling the strain environment of the colour centre, we tune its electronic levels to probe, control, and eventually suppress the interaction of its spin with the thermal bath. Strain control provides both large tunability of the optical transitions and significantly improved spin coherence. Finally, our findings indicate the possibility to achieve strong coupling between the silicon-vacancy spin and single phonons, which can lead to the realisation of phonon-mediated quantum gates and nonlinear quantum phononics.

Full counting statistics of quantum dot resonance fluorescence.

The electronic energy levels and optical transitions of a semiconductor quantum dot are subject to dynamics within the solid-state environment. In particular, fluctuating electric fields due to nearby charge traps or other quantum dots shift the transition frequencies via the Stark effect. The environment dynamics are mapped directly onto the fluorescence under resonant excitation and diminish the prospects of quantum dots as sources of indistinguishable photons in optical quantum computing. Here, we present an analysis of resonance fluorescence fluctuations based on photon counting statistics which captures the underlying time-averaged electric field fluctuations of the local environment. The measurement protocol avoids dynamic feedback on the electric environment and the dynamics of the quantum dot's nuclear spin bath by virtue of its resonant nature and by keeping experimental control parameters such as excitation frequency and external fields constant throughout. The method introduced here is experimentally undemanding.

Contamination rates and antimicrobial resistance in Enterococcus spp., Escherichia coli, and Salmonella isolated from "no antibiotics added"-labeled chicken products.

In the United States, products from chickens that were not administered antimicrobial medications during growout can contain labels stating "no antibiotics added." Here we compared microbial profiles of chicken products labeled as coming from birds raised without antimicrobial medications (N=201; NON) with chicken products carrying conventional labels (N=201; CONV). There were no differences in percentages of samples positive for Enterococcus spp. (CONV: 17.4%; NON: 21.3%) or Escherichia coli (CONV: 25.9%; NON: 22.3%). The number of samples positive for Salmonella was low in both groups, but statistically higher in the NON samples (5.0%) versus CONV samples (1.5%; p<0.05). Conversely, CONV samples contained higher concentrations of coliforms (CONV: 3.0 log(10)CFU/mL; NON: 2.5 log(10)CFU/mL; p<0.05). E. coli (N=190) and Enterococcus spp. isolates (N=113) were tested for resistance to common antimicrobials. E. coli isolates from CONV samples were more frequently resistant to at least one antimicrobial (CONV: 61.3%; NON: 41.2%; p<0.05). Enterococcus spp. isolates from both groups were equally likely to be resistant to at least one antimicrobial, but Enterococcus spp. isolates from CONV samples were more likely to be resistant to erythromycin, kanamycin, and gentamicin (p<0.05). Taken together, these data suggest that NON samples may more frequently carry Salmonella; however, E. coli and Enterococcus spp. found on CONV are more likely to be resistant to some antimicrobials.