Predicting the Risk of Injuries Through Assessments of Asymmetric Lower Limb Functional Performance: A Prospective Study of 415 Youth Taekwondo Athletes.

Background: The impact of interlimb asymmetries on sport injuries is unclear because of inconsistent findings, and there is a lack of research on youth athletes and the sport of taekwondo. Purpose: To examine the effects of functional interlimb asymmetries on noncontact lower limb injuries in youth athletes. Study Design: Cohort study; Level of evidence, 2. Methods: A total of 415 taekwondo athletes (318 boys and 97 girls) aged 6 to 17 years underwent baseline testing to determine interlimb asymmetries through the single-leg countermovement jump (CMJ), hop, and triple hop tests as well as the Star Excursion Balance Test. The athletes were then evaluated for 12 months to observe the occurrence of noncontact lower limb injuries. Results: During the study, 98 athletes (70 boys and 28 girls) sustained at least 1 noncontact lower limb injury. Athletes with higher interlimb asymmetries in single-leg CMJ height showed a significantly increased risk of noncontact lower limb injuries (boys: odds ratio [OR], 1.053 [95% CI, 1.027-1.080], P < .001; girls: OR, 1.070 [95% CI, 1.016-1.128], P = .011). Asymmetry in single-leg CMJ height of ≥15.28% was found to be the cutoff point for predicting noncontact lower limb injuries in boys (OR, 4.652 [95% CI, 2.577-8.398]; P < .001). Conclusion: This study highlights the utility of interlimb asymmetries in unilateral jump performance as a tool for assessing the risk of noncontact lower limb injuries in youth taekwondo athletes of both sexes. A proper evaluation of interlimb asymmetries may improve prevention strategies for youth athletes.

SNAr Regioselectivity Predictions: Machine Learning Triggering DFT Reaction Modeling through Statistical Threshold.

Fast and accurate prospective predictions of regioselectivity can significantly reduce the time and resources spent on unproductive transformations in the pharmaceutical industry. Density functional theory (DFT) reaction modeling through transition state theory (TST) and machine learning (ML) methods has been widely used to predict reaction outcomes such as selectivity. However, TST reaction modeling and ML methods are either time-consuming or data-dependent. Herein, we introduce a prototype seamlessly bridging ML and TST modeling by triggering resource-intensive but much less domain-sensitive DFT calculations only on less confident ML predictions. The proposed workflow was trained and tested on both the Pfizer internal dataset and the USPTO public dataset to predict regioselectivity for SNAr reactions. Our method is accurate and fast, which achieves 96.3 and 94.7% accuracy in predicting the correct major product on Pfizer and USPTO datasets, respectively, in a fraction of conventional TST computing time.

The Single-Component Flavin Reductase/Flavin-Dependent Halogenase AetF is a Versatile Catalyst for Selective Bromination and Iodination of Arenes and Olefins.

Flavin-dependent halogenases (FDHs) natively catalyze selective halogenation of electron rich aromatic and enolate groups. Nearly all FDHs reported to date require a separate flavin reductase to supply them with FADH2 , which complicates biocatalysis applications. In this study, we establish that the single component flavin reductase/flavin dependent halogenase AetF catalyzes halogenation of a diverse set of substrates using a commercially available glucose dehydrogenase to drive its halogenase activity. High site selectivity, activity on relatively unactivated substrates, and high enantioselectivity for atroposelective bromination and bromolactonization was demonstrated. Site-selective iodination and enantioselective cycloiodoetherification was also possible using AetF. The substrate and reaction scope of AetF suggest that it has the potential to greatly improve the utility of biocatalytic halogenation.

Associations between Sleep Characteristics and Cardiovascular Risk Factors in Adolescents Living with Type 1 Diabetes.

Adolescents living with type 1 diabetes (T1D) have an increased risk of developing cardiovascular disease. Sleep patterns have physiological and behavioral impacts on diabetes outcomes. This study aimed to investigate the associations between sleep patterns and CVD risk factors in adolescents living with T1D and their peers living without T1D. This cross-sectional study assessed CVD risk factors and sleep characteristics (and their associations) in adolescents, aged 12-18 years, living with T1D (n = 48) and their peers (n = 19) without T1D. Outcomes included blood pressure, lipid profiles, and sleep characteristics (accelerometry). Statistical differences between groups were determined with chi-square or independent samples t-tests. The associations between sleep characteristics and CVD risk factors were assessed with multivariate linear regression analyses. We found no significant differences between the two groups in terms of sleep duration, efficiency, sleep onset and offset, and frequency of awakenings, and there were associations between sleep efficiency and LDL-C (ß = -0.045, p = 0.018, model R2 = 0.230) and triglycerides (ß = -0.027, p = 0.012, model R2 = 0.222) after adjusting confounders (diabetes status, sex, age, pubertal stage) in all participants. In conclusion, adolescents with T1D and without T1D sleep less than the recommended eight hours per night. The associations between sleep efficiency and LDL-C and triglycerides are independent of sleep duration, regardless of sex, age, and pubertal stage.

Association between Inter-Limb Asymmetries in Lower-Limb Functional Performance and Sport Injury: A Systematic Review of Prospective Cohort Studies.

BACKGROUND: Inter-limb asymmetry in lower-limb functional performance has been associated with increased risk of sport injury; however, findings are not always consistent. PURPOSE: To conduct a systematic review on whether inter-limb asymmetry in lower-limb functional performance can predict sport injury. METHODS: Four electronic databases (MEDLINE, EMBASE, Web of Science, and SportDiscus) were systematically searched for prospective cohort studies reporting the association between inter-limb asymmetry in lower-limb functional performance and sport injury. RESULTS: A total of 28 prospective cohort studies were included in the analyses. Collectively, the findings were highly inconsistent, and a clear statement on the association between each asymmetry and sport injury was difficult. CONCLUSIONS: Highly inconsistent findings make it difficult to create clear recommendations on the relationship between the inter-limb asymmetry in lower-limb functional performance (power, muscle flexibility, and dynamic balance) and sport injury. The influence of potential factors (selection of tests/parameters, participant characteristics, definition of injury, and ways of calculating asymmetry) should be considered when using previous findings.

Real-time prediction of 1H and 13C chemical shifts with DFT accuracy using a 3D graph neural network.

Nuclear magnetic resonance (NMR) is one of the primary techniques used to elucidate the chemical structure, bonding, stereochemistry, and conformation of organic compounds. The distinct chemical shifts in an NMR spectrum depend upon each atom's local chemical environment and are influenced by both through-bond and through-space interactions with other atoms and functional groups. The in silico prediction of NMR chemical shifts using quantum mechanical (QM) calculations is now commonplace in aiding organic structural assignment since spectra can be computed for several candidate structures and then compared with experimental values to find the best possible match. However, the computational demands of calculating multiple structural- and stereo-isomers, each of which may typically exist as an ensemble of rapidly-interconverting conformations, are expensive. Additionally, the QM predictions themselves may lack sufficient accuracy to identify a correct structure. In this work, we address both of these shortcomings by developing a rapid machine learning (ML) protocol to predict 1H and 13C chemical shifts through an efficient graph neural network (GNN) using 3D structures as input. Transfer learning with experimental data is used to improve the final prediction accuracy of a model trained using QM calculations. When tested on the CHESHIRE dataset, the proposed model predicts observed 13C chemical shifts with comparable accuracy to the best-performing DFT functionals (1.5 ppm) in around 1/6000 of the CPU time. An automated prediction webserver and graphical interface are accessible online at http://nova.chem.colostate.edu/cascade/. We further demonstrate the model in three applications: first, we use the model to decide the correct organic structure from candidates through experimental spectra, including complex stereoisomers; second, we automatically detect and revise incorrect chemical shift assignments in a popular NMR database, the NMRShiftDB; and third, we use NMR chemical shifts as descriptors for determination of the sites of electrophilic aromatic substitution.

Risk Factors for Non-Contact Lower-Limb Injury: A Retrospective Survey in Pediatric-Age Athletes.

BACKGROUND: Risk factors for non-contact lower-limb injury in pediatric-age athletes and the effects of lateral dominance in sport (laterally vs. non-laterally dominant sports) on injury have not been investigated. PURPOSE: To identify risk factors for non-contact lower-limb injury in pediatric-age athletes. METHODS: Parents and/or legal guardians of 2269 athletes aged between 6-17 years were recruited. Each participant completed an online questionnaire that contained 10 questions about the athlete's training and non-contact lower-limb injury in the preceding 12 months. RESULTS: The multivariate logistic regression model determined that lateral dominance in sport (adjusted OR (laterally vs. non-laterally dominant sports), 1.38; 95% CI, 1.10-1.75; p = 0.006), leg preference (adjusted OR (right vs. left-leg preference), 0.71; 95% CI, 0.53-0.95; p = 0.023), increased age (adjusted OR, 1.21; 95% CI, 1.16-1.26; p = 0.000), training intensity (adjusted OR, 1.77; 95% CI, 1.43-2.19; p = 0.000), and training frequency (adjusted OR, 1.36; 95% CI, 1.25-1.48; p = 0.000) were significantly associated with non-contact lower-limb injury in pediatric-age athletes. Length of training (p = 0.396) and sex (p = 0.310) were not associated with a non-contact lower-limb injury. CONCLUSIONS: Specializing in laterally dominant sports, left-leg preference, increase in age, training intensity, and training frequency indicated an increased risk of non-contact lower-limb injury in pediatric-age athletes. Future research should take into account exposure time and previous injury.

Predicting Infrared Spectra with Message Passing Neural Networks.

Infrared (IR) spectroscopy remains an important tool for chemical characterization and identification. Chemprop-IR has been developed as a software package for the prediction of IR spectra through the use of machine learning. This work serves the dual purpose of providing a trained general-purpose model for the prediction of IR spectra with ease and providing the Chemprop-IR software framework for the training of new models. In Chemprop-IR, molecules are encoded using a directed message passing neural network, allowing for molecule latent representations to be learned and optimized for the task of spectral predictions. Model training incorporates spectra metrics and normalization techniques that offer better performance with spectral predictions than standard practice in regression models. The model makes use of pretraining using quantum chemistry calculations and ensembling of multiple submodels to improve generalizability and performance. The spectral predictions that result are of high quality, showing capability to capture the extreme diversity of spectral forms over chemical space and represent complex peak structures.

Association between physical activity level and cardiovascular risk factors in adolescents living with type 1 diabetes mellitus: a cross-sectional study.

BACKGROUND: Type 1 diabetes mellitus (T1DM) is associated with an increased risk for cardiovascular disease (CVD) related morbidity and premature mortality. Regular physical activity plays an important role in the primary and secondary prevention of CVD, improving overall health and wellbeing. Previous observational studies have examined the associations between self-reported physical activity and CVD risk factors in largely adult Caucasian populations. However, limited work has evaluated the relationship between objectively measured physical activity and CVD risk factors in other ethnicities, particularly Chinese youth living with T1DM. METHODS: This cross-sectional study assessed CVD risk factors, physical activity, and aerobic fitness (and their associations) in Chinese youth living with T1DM (n = 48) and peers (n = 19) without T1DM. Primary outcomes included blood pressure, lipid profiles, and physical activity (accelerometry). Statistical differences between groups were determined with chi-square, independent-samples t-tests, or analysis of covariance. The associations between aerobic fitness, daily physical activity variables, and CVD risk factors were assessed with univariate and multivariate linear regression analyses. RESULTS: Results were summarized using means and standard deviation (SD) for normally distributed variables and medians and 25-75th quartile for non-normally distributed variables. In comparison to peers without diabetes, youth living with T1DM showed higher levels of total cholesterol (3.14 ± 0.67 vs. 4.03 ± 0.81 mmol·L-1, p = 0.001), low-density lipoprotein cholesterol (1.74 ± 0.38 vs. 2.31 ± 0.72 mmol·L-1, p = 0.005), and triglycerides (0.60 ± 0.40 vs. 0.89 ± 0.31 mmol·L-1 p = 0.012), and lower maximal oxygen power (44.43 ± 8.29 vs. 35.48 ± 8.72 mL·kg-1·min-1, p = 0.003), total physical activity counts (451.01 ± 133.52 vs. 346.87 ± 101.97 counts·min-1, p = 0.004), metabolic equivalents (METs) (2.41 ± 0.60 vs. 2.09 ± 0.41 METs, p = 0.033), moderate-to-vigorous intensity physical activity [MVPA: 89.57 (61.00-124.14) vs (53.19 (35.68-63.16) min, p = 0.001], and the percentage of time spent in MVPA [11.91 (7.74-16.22) vs 8.56 (6.18-10.12) %, p = 0.038]. The level of high-density lipoprotein cholesterol was positively associated with METs (ß = 0.29, p = 0.030, model R2 = 0.168), and the level of triglycerides was negatively associated with physical activity counts (ß = - 0.001, p = 0.018, model R2 = 0.205) and METs (ß = - 0.359, p = 0.015, model R2 = 0.208), and positively associated with time spent in sedentary behaviour (ß = 0.002, p = 0.041, model R2 = 0.156) in persons living with T1DM. CONCLUSIONS: Chinese youth with T1DM, despite their young age and short duration of diabetes, present early signs of CVD risk, as well as low physical activity levels and cardiorespiratory fitness compared to apparently healthy peers without diabetes. Regular physical activity is associated with a beneficial cardiovascular profile in T1DM, including improvements in lipid profile. Thus, physical activity participation should be widely promoted in youth living with T1DM.

The effect of fatigue on asymmetry between lower limbs in functional performances in elite child taekwondo athletes.

BACKGROUND: Inter-limb asymmetry above a certain threshold in functional performance indicates increased injury risk in sports. Fatigue has been found to increase bilateral asymmetry in lower-limb jumping performance among high-school and adult athletes, whereas this impact has not been examined in child athletes. This study aimed to examine the effect of fatigue on inter-limb asymmetry in functional performances in elite Taekwondo athletes aged between 9 and 11 years. METHODS: Performance of single-leg jumps, Star Excursion Balance Test (SEBT), and muscle (hamstring and gastrocnemius) flexibility were measured for 13 elite male child Taekwondo athletes (aged 9.85 ± 0.80 years) at both the rested and fatigued states to examine the inter-limb asymmetry. A two-way repeated measures ANOVA was conducted to examine for difference and the interaction between limb (dominant, non-dominant leg) and state (rested, fatigued state) for each test. Paired t test or Wilcoxon signed-rank test was used to compare the asymmetry magnitude at the rested vs. fatigued state for each test, and the variation of performance post fatigue in the dominant vs. non-dominant leg when appropriate. RESULTS: The inter-limb asymmetry in triple-hop distance significantly (p = 0.046) increased with fatigue, whereas the asymmetry significantly (p = 0.004) decreased with fatigue in anterior (ANT) reach distance in SEBT. A significant (p = 0.027) limb by state interaction was shown for posterolateral (PL) reach distance in SEBT, wherein a significant (p = 0.005) bilateral difference was only shown at the rested state. The PL reach distance showed a significantly greater decrease (p = 0.028) post fatigue when using the dominant leg for support compared to using the non-dominant leg. CONCLUSIONS: Fatigue significantly impacts inter-limb asymmetry in jump performances and dynamic balance for child athletes, while the variation of inter-limb asymmetry post fatigue may be different across tests. For the purpose of injury prevention, practitioners should consider assessing the inter-limb asymmetry for children at both the rested and fatigued state and be mindful of the fatigue response of each leg in functional tests.

Bilateral difference between lower limbs in children practicing laterally dominant vs. non-laterally dominant sports.

Bilateral asymmetry in lower-limb power and dynamic balance has been associated with increased risk of sport injury, whereas there is a lack of research examining this asymmetry for child athletes. Twenty-eight fencers (19 boys and 9 girls, aged 9.71 ± 1.08 years) and 28 Taekwondo athletes (19 boys and 9 girls, aged 9.71 ± 1.08 years) were examined on the single-leg jump and Star Excursion Balance Test (SEBT) performance. A mixed model design ANOVA (2 [Sex: Boys, Girls] × 2 [Sport group: Fencing, Taekwondo] × 2 [Limb: Dominant, Non-dominant Leg]) was conducted to examine for difference for each test. There was a significant main effect of limb on hop and triple hop distance (p < 0.05). A significant limb by sex interaction (p = 0.000) was shown for the single-leg countermovement jump (CMJ) performance, wherein a bilateral difference was only shown in boys. In SEBT, a main effect (p = 0.007) of limb was identified for posterolateral (PL) reach distance. A limb by sex interaction (p = 0.009) was also shown for posteromedial (PM) reach distance, wherein a bilateral difference was only shown in girls. These findings suggest that child athletes in both laterally dominant and non-laterally dominant sports showed inter-limb asymmetry of leg power and dynamic balance. Sex should be an important consideration when evaluating bilateral difference of leg power and dynamic balance for child athletes.

Quantum chemical calculations for over 200,000 organic radical species and 40,000 associated closed-shell molecules.

The stabilities of radicals play a central role in determining the thermodynamics and kinetics of many reactions in organic chemistry. In this data descriptor, we provide consistent and validated quantum chemical calculations for over 200,000 organic radical species and 40,000 associated closed-shell molecules containing C, H, N and O atoms. These data consist of optimized 3D geometries, enthalpies, Gibbs free energy, vibrational frequencies, Mulliken charges and spin densities calculated at the M06-2X/def2-TZVP level of theory, which was previously found to have a favorable trade-off between experimental accuracy and computational efficiency. We expect this data to be useful in the further development of machine learning techniques to predict reaction pathways, bond strengths, and other phenomena closely related to organic radical chemistry.

Publisher Correction: Prediction of organic homolytic bond dissociation enthalpies at near chemical accuracy with sub-second computational cost.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Prediction of organic homolytic bond dissociation enthalpies at near chemical accuracy with sub-second computational cost.

Bond dissociation enthalpies (BDEs) of organic molecules play a fundamental role in determining chemical reactivity and selectivity. However, BDE computations at sufficiently high levels of quantum mechanical theory require substantial computing resources. In this paper, we develop a machine learning model capable of accurately predicting BDEs for organic molecules in a fraction of a second. We perform automated density functional theory (DFT) calculations at the M06-2X/def2-TZVP level of theory for 42,577 small organic molecules, resulting in 290,664 BDEs. A graph neural network trained on a subset of these results achieves a mean absolute error of 0.58 kcal mol-1 (vs DFT) for BDEs of unseen molecules. We further demonstrate the model on two applications: first, we rapidly and accurately predict major sites of hydrogen abstraction in the metabolism of drug-like molecules, and second, we determine the dominant molecular fragmentation pathways during soot formation.

Anti-depression effects of ketogenic diet are mediated via the restoration of microglial activation and neuronal excitability in the lateral habenula.

Depression is a severe neuropsychiatric disorder, of which the underlying pathological mechanisms remain unclear. The ketogenic diet (KD) has been reported to exhibit preventative effects on depressive-like behaviors in rodents. However, the therapeutic effects of KD on depressive-like behaviors have not been illustrated thus far. Here, we found that KD treatment dramatically ameliorated depressive-like behaviors in both repeated social defeat stress (R-SDS) and lipopolysaccharide (LPS) models, indicating the potential therapeutic effects of KD on depression. Our electrophysiological studies further showed that neuronal excitability was increased in the lateral habenula (LHb) of mice exposed to R-SDS or LPS, which can be reversed in the presence of KD treatment. Moreover, R-SDS and LPS were also found to induce robust microglial inflammatory activation in the LHb. Importantly, these phenotypes were rescued in mice fed with KD. In addition, we found that the protein level of innate immune receptor Trem2 in the LHb was significantly decreased in depression models. Specific knockdown of Trem2 in LHb microglia induced depressive-like behaviors, increased neuronal excitability as well as robust microglial inflammatory activation. Altogether, we demonstrated the therapeutic effects of KD on depressive-like behaviors, which are probably mediated via the restoration of microglial inflammatory activation and neuronal excitability. Besides, we also proposed an unrecognized function of Trem2 in the LHb for depression. Our study sheds light on the pathogenesis of depression and thereby offers a potential therapeutic intervention.

[The value of P16/ki67 double labeling, HPV E6/E7 mRNA detection and their combined application in cytological shunt diagnosis of low-grade squamous intraepithelial lesions].

Objective To explore the value of double labeling of P16/ki67, E6/E7 mRNA of human papillomavirus (HPV) and combined detection in shunt diagnosis of low-grade squamous intraepithelial lesions (LSIL) by thin-layer cervical cytology (TCT). Methods The study enrolled 239 patients who underwent colposcopy and biopsy within 4 weeks after primary TCT diagnosis. The remaining cytological samples were double-labeled with P16/ki67 immunocytochemical staining and the HPV E6/E7 mRNA was detected by Panther automatic HPV E6/E7 mRNA detection system. Using SPSS22.0 software, the positive rates of P16/ki67 double-labeling, HPV E6/E7 mRNA and combined detection were analyzed in different cervical lesions, and the positive rates in the same cervical lesions were compared horizontally to evaluate the efficiency of double labeling of P16/ki67, HPV E6/E7 and combined detection in the diagnosis of high-grade squamous intraepithelial lesions (HSIL) and above lesions. Results The diagnostic results of HE staining for the 239 cases of LSIL were 71 cases of chronic cervicitis (29.71%), 143 cases of LSIL (59.83%), 22 cases of HSIL (9.20%) and 3 cases of cervical cancer (1.26%). There were 46 cases of P16+ki67+ lesions (19.25%), 41 cases of ki67+P16- lesions (17.15%), 33 cases of ki67-P16+ lesions (13.81%) and 119 cases of P16-ki67- lesions (49.79%). The positive rates of P16/ki67 double-labeling, HPV E6/E7 mRNA and combined detection increased with the severity of cervical lesions. The positive rate of combined detection was the highest in the HSIL lesions, which was higher than that of P16/ki67 double-labeling and HPV E6/E7 mRNA detection. The sensitivity of combined detection was higher than that of P16/ki67 double-labeling and HPV E6/E7 mRNA detection. The Youden index of joint detection was 0.7850. Conclusion The combined detection of P16/ki67 double labeling, HPV E6/E7 mRNA and HPV E6/E7 mRNA had a certain clinical value in the management of cell LSIL shunt diagnosis. The combined detection significantly improved the sensitivity and Youden index of HSIL and above lesions, while maintaining a high specificity and coincidence rate.

Social Company by a Receptive Mating Partner Facilitates Fear Extinction.

Fear extinction remains an unresolved challenge for behavioral exposure therapy in patients with post-traumatic stress disorder (PTSD). Previous reports have suggested that social support from either familiar or unfamiliar same-sex partners is beneficial to attenuating fear responses during fear extinction and renewal. Despite that, few studies have examined the effects of social support in advance on fear extinction and/or retrieval. It is also not clear whether social company by a receptive mating partner in advance facilitates fear extinction. In the present study, we address these questions by introducing a co-housing method, where fear-conditioned male mice are co-housed with or without a receptive mating partner prior to fear extinction. We found that while co-housing with an ovariectomized female mouse showed little effect on fear extinction or retrieval, social company by a receptive mating partner in advance dramatically facilitates fear extinction. In addition, the number of cFos-positive neurons in the basolateral amygdala (BLA) were also found to be reduced in male mice accompanied with receptive mating partner in response to fear extinction and retrieval, indicating diminished neuronal activation. Electrophysiological studies further showed that the excitability of excitatory neurons in BLA was decreased, which is probably due to the attenuated basal level of excitatory synaptic transmission. Together, our observations demonstrate an effect of social company by a receptive mating partner can facilitate fear extinction and afford a possible cellular mechanism.

Erbin in Amygdala Parvalbumin-Positive Neurons Modulates Anxiety-like Behaviors.

BACKGROUND: Anxiety disorders are the most common psychiatric diseases, affecting 28% of people worldwide within their lifetime. The excitation-inhibition imbalance in the amygdala is thought to be an underlying pathological mechanism; however, the cellular and molecular control of amygdala excitation-inhibition balance is largely unknown. METHODS: By using mice expressing chemogenetic activator or inhibitor channel in amygdala parvalbumin (PV) neurons, Erbin mutant mice, and mice with Erbin specifically knocked down in amygdala PV neurons, we systematically investigated the role of amygdala PV neurons and Erbin expressed therein in the pathogenesis of anxiety disorders using the combined approaches of immunohistochemistry, electrophysiology, and behavior. RESULTS: In naïve mice, chemogenetic inhibition of PV neurons produced anxiogenic effects, suggesting an essential role in the regulation of anxiety. In stressed mice with anxiety, excitatory postsynaptic responses on amygdala PV neurons were selectively diminished, accompanied by a decreased expression of Erbin specifically in amygdala PV neurons. Remarkably, both Erbin mutant mice and amygdala PV-specific Erbin knockdown mice exhibited impaired excitatory postsynaptic responses on amygdala PV neurons and increased anxiety-like behaviors. Furthermore, chemogenetic activation of amygdala PV neurons normalized anxiety behaviors in amygdala PV-specific Erbin knockdown mice and stressed mice. CONCLUSIONS: Together, these results demonstrate that Erbin in PV neurons is critical for maintaining the excitation-inhibition balance in the amygdala and reveal a novel pathophysiological mechanism for anxiety disorders.

Evaluating and clustering retrosynthesis pathways with learned strategy.

With recent advances in the computer-aided synthesis planning (CASP) powered by data science and machine learning, modern CASP programs can rapidly identify thousands of potential pathways for a given target molecule. However, the lack of a holistic pathway evaluation mechanism makes it challenging to systematically prioritize strategic pathways except for using some simple heuristics. Herein, we introduce a data-driven approach to evaluate the relative strategic levels of retrosynthesis pathways using a dynamic tree-structured long short-term memory (tree-LSTM) model. We first curated a retrosynthesis pathway database, containing 238k patent-extracted pathways along with ∼55 M artificial pathways generated from an open-source CASP program, ASKCOS. The tree-LSTM model was trained to differentiate patent-extracted and artificial pathways with the same target molecule in order to learn the strategic relationship among single-step reactions within the patent-extracted pathways. The model achieved a top-1 ranking accuracy of 79.1% to recognize patent-extracted pathways. In addition, the trained tree-LSTM model learned to encode pathway-level information into a representative latent vector, which can facilitate clustering similar pathways to help illustrate strategically diverse pathways generated from CASP programs.

Regio-selectivity prediction with a machine-learned reaction representation and on-the-fly quantum mechanical descriptors.

Accurate and rapid evaluation of whether substrates can undergo the desired the transformation is crucial and challenging for both human knowledge and computer predictions. Despite the potential of machine learning in predicting chemical reactivity such as selectivity, popular feature engineering and learning methods are either time-consuming or data-hungry. We introduce a new method that combines machine-learned reaction representation with selected quantum mechanical descriptors to predict regio-selectivity in general substitution reactions. We construct a reactivity descriptor database based on ab initio calculations of 130k organic molecules, and train a multi-task constrained model to calculate demanded descriptors on-the-fly. The proposed platform enhances the inter/extra-polated performance for regio-selectivity predictions and enables learning from small datasets with just hundreds of examples. Furthermore, the proposed protocol is demonstrated to be generally applicable to a diverse range of chemical spaces. For three general types of substitution reactions (aromatic C-H functionalization, aromatic C-X substitution, and other substitution reactions) curated from a commercial database, the fusion model achieves 89.7%, 96.7%, and 97.2% top-1 accuracy in predicting the major outcome, respectively, each using 5000 training reactions. Using predicted descriptors, the fusion model is end-to-end, and requires approximately only 70 ms per reaction to predict the selectivity from reaction SMILES strings.