DeepP450: Predicting Human P450 Activities of Small Molecules by Integrating Pretrained Protein Language Model and Molecular Representation.

Cytochrome P450 enzymes (CYPs) play a crucial role in Phase I drug metabolism in the human body, and CYP activity toward compounds can significantly affect druggability, making early prediction of CYP activity and substrate identification essential for therapeutic development. Here, we established a deep learning model for assessing potential CYP substrates, DeepP450, by fine-tuning protein and molecule pretrained models through feature integration with cross-attention and self-attention layers. This model exhibited high prediction accuracy (0.92) on the test set, with area under the receiver operating characteristic curve (AUROC) values ranging from 0.89 to 0.98 in substrate/nonsubstrate predictions across the nine major human CYPs, surpassing current benchmarks for CYP activity prediction. Notably, DeepP450 uses only one model to predict substrates/nonsubstrates for any of the nine CYPs and exhibits certain generalizability on novel compounds and different categories of human CYPs, which could greatly facilitate early stage drug design by avoiding CYP-reactive compounds.

EvoAI enables extreme compression and reconstruction of the protein sequence space.

Designing proteins with improved functions requires a deep understanding of how sequence and function are related, a vast space that is hard to explore. The ability to efficiently compress this space by identifying functionally important features is extremely valuable. Here, we first establish a method called EvoScan to comprehensively segment and scan the high-fitness sequence space to obtain anchor points that capture its essential features, especially in high dimensions. Our approach is compatible with any biomolecular function that can be coupled to a transcriptional output. We then develop deep learning and large language models to accurately reconstruct the space from these anchors, allowing computational prediction of novel, highly fit sequences without prior homology-derived or structural information. We apply this hybrid experimental-computational method, which we call EvoAI, to a repressor protein and find that only 82 anchors are sufficient to compress the high-fitness sequence space with a compression ratio of 1048. The extreme compressibility of the space informs both applied biomolecular design and understanding of natural evolution.

RTEL1 gene polymorphisms and neuroblastoma risk in Chinese children.

BACKGROUND: Neuroblastoma, a neuroendocrine tumor originating from the sympathetic ganglia, is one of the most common malignancies in childhood. RTEL1 is critical in many fundamental cellular processes, such as DNA replication, DNA damage repair, genomic integrity, and telomere stability. Single nucleotide polymorphisms (SNPs) in the RTEL1 gene have been reported to confer susceptibility to multiple cancers, but their contributing roles in neuroblastoma remain unclear. METHODS: We conducted a study on 402 neuroblastoma cases and 473 controls to assess the association between four RTEL1 SNPs (rs3761124 T>C, rs3848672 T>C, rs3208008 A>C and rs2297441 G>A) and neuroblastoma susceptibility. RESULTS: Our results show that rs3848672 T>C is significantly associated with an increased risk of neuroblastoma [CC vs. TT/TC: adjusted odds ratio (OR)=1.39, 95% confidence interval (CI)=1.02-1.90, P=0.038]. The stratified analysis further indicated that boy carriers of the rs3848672 CC genotype had a higher risk of neuroblastoma, and all carriers had an increased risk of developing neuroblastoma of mediastinum origin. Moreover, the rs2297441 AA genotype increased neuroblastoma risk in girls and predisposed children to neuroblastoma arising from retroperitoneal. CONCLUSION: Our study indicated that the rs3848672 CC and rs2297441 AA genotypes of the RTEL1 gene are significantly associated with an increased risk of neuroblastoma in Chinese children in a gender- and site-specific manner.

Machine learning strategy on activation energy of environmental heterogeneous reactions and its application to atmospheric formation of typical montmorillonite-bound phenoxy radicals.

Heterogeneous transformation of organic pollutants into more toxic chemicals poses substantial health risks to humans. Activation energy is an important indicator that help us to understand transformation efficacy of environmental interfacial reactions. However, the determination of activation energies for large numbers of pollutants using either the experimental or high-accuracy theoretical methods is expensive and time-consuming. Alternatively, the machine learning (ML) method shows the strength in predictive performance. In this study, using the formation of a typical montmorillonite-bound phenoxy radical as an example, a generalized ML framework RAPID was proposed for activation energy prediction of environmental interfacial reactions. Accordingly, an explainable ML model was developed to predict the activation energy via easily accessible properties of the cations and organics. The model developed by decision tree (DT) performed best with the lowest root-mean-squared error (RMSE = 0.22) and the highest coefficient of determination values (R2 score = 0.93), the underlying logic of which was well understood by combining model visualization and SHapley Additive exPlanations (SHAP) analysis. The performance and interpretability of the established model suggest that activation energies can be predicted by the well-designed ML strategy, and this would allow us to predict more heterogeneous transformation reactions in the environmental field.

Chemical investigations and cytotoxic effects of metabolites from Antrodia camphorata against human hepatocellular carcinoma cells.

Antrodia camphorata is used as a medicinal fungus in Taiwan to treat fatigue, food intoxication, and enhance liver function. Here we identified fermented metabolic components from the mycelium of A. camphorata KH37 and explored their anti-hepatoma potentials with study models of human hepatoblastoma cell lines. Bioassay-guided fractionation of the solid fermentation powder of A. camphorata KH37 led to the isolation of one new quinonol, antroquinonol Z (1), and nine known compounds (2-10). Treatment with 10 µM antrocamols LT1 (2) or LT3 (3) reduced cell viability of HepG2 and Huh-7 cells to about 60% in 48 hours. Antroquinonol Z (1) exhibited mild cytotoxicity against Huh-7 cells in 48 and 72 hours. Interestingly, two fractions showed cytotoxicity in HepG2 and Huh-7 cells, even better than compounds isolated from these fractions. The significant cytotoxicity of partially purified samples from A. camphorata KH37 exhibited a potential for developing alternative or complementary therapeutics against hepatoma.

The formation of PBDFs from the ortho-disubstituted phenol precursors: A comprehensive theoretical study on the PBDD/Fs formation from 2,4,6-tribromophenol.

Bromophenols are known as direct precursors of the notorious polybrominated dibenzo-p-dioxin/dibenzofurans (PBDD/Fs). There is a long-held viewpoint that only the more toxic dioxin-type products could be formed from the ortho-disubstituted phenols, totally contrary to the experimental observations that both PBDDs and PBDFs are generated. To tackle the issue, the gaseous formation mechanism of PBDD/Fs from 2,4,6-tribromophenol (TBP), a typical ortho-disubstituted phenol, was investigated in this study. Firstly, the reactions between TBP and the active H radical produce three key radical species including the bromophenoxyl radical, the substituted phenyl radical and phenoxyl diradical. The self- and cross-combinations of these radical species and TBP yield not only the dioxin-type products 1,3,6,8-TeBDD and 1,3,7,9-TeBDD, but also the brominated dibenzofurans 1,3,6,8-TeBDF and 2,4,6,8-TeBDF. Notably, the reactions involving the phenyl C sites in the substituted phenyl and phenoxyl diradicals are demonstrated to be both thermodynamically and kinetically more favorable than those involving the bromophenoxyl radical and the TBP molecule. Most importantly, the findings of the present work are of great importance as it provides feasible pathways to form less toxic dibenzofuran-type products from the ortho-disubstituted phenols. These results will improve the understanding of the PBDD/Fs formation mechanism from phenol precursors.

Interfacial formation of environmentally persistent free radicals-A theoretical investigation on pentachlorophenol activation on montmorillonite in PM2.5.

Environmentally persistent free radicals (EPFRs) in atmospheric fine particulate matters (PM2.5) possess high bioactivity and result in severe health problems. The facile transformation of aromatic pollutants into EPFRs on montmorillonite (MMT), an important solid component in PM2.5, is an activation of air pollutants into more toxic chemical species and also attributes to the secondary source of EPFRs in PM2.5. In this study, the interfacial reactions of pentachlorophenol (PCP), a typical EPFR precursor in air pollution, on the Fe(III)-, Ca- and Na-MMT surfaces have been explored by the density functional theory (DFT) calculations using the periodic slab models. The PCP molecule is found to be exothermically adsorbed on the three MMT surfaces. Moreover, significant charge transfer from PCP to Fe takes place and finally leads to the surface-bound phenoxyl radical formation on the Fe(III)-MMT surface since the half-filled 3d orbital of Fe3+ in Fe(III)-MMT could act as electron acceptor allowing the electron transferring from the 2p orbital of the phenolic O in PCP to Fe ion. However, similar charge transfer is not found in the Ca- and Na-MMTs, and the PCP transformation reaction is hindered on the Ca- and Na-MMT surfaces. Namely, the PCP activation to the corresponding EPFRs is impossible on the Ca-MMT and Na-MMT surfaces, while the catalytically active Fe(III)-MMT in PM2.5 can transform the chlorinated phenols into more toxic phenoxy-type EPFRs at ambient temperatures. Accordingly, more attention should be paid on the effect of MMT with catalytical capacity on the toxicity of PM2.5.

Theoretical study on the formation mechanism of polychlorinated dibenzothiophenes/thianthrenes from 2-chlorothiophenol molecules.

Homogeneous formation of polychlorinated dibenzothiophenes/thianthrenes (PCDT/TAs), sulfurated compounds analogous to polychlorinated dibenzo-p-dioxin/dibenzofurans (PCDD/Fs), has been well-documented to occur via radical-radical coupling reactions from chlorinated thiophenol precursors. However, the current understanding of the formation mechanism of PCDT/TAs is exclusively limited to the inherent point of view that chlorothiophenoxy radicals act as the only required intermediates for PCDT/TAs. This study investigates reaction pathways for the formation of PCDT/TAs involving two new types of radical species, i.e., substituted phenyl radicals and substituted thiophenoxyl diradicals. Taking 2-chlorothiophenol (2-CTP) as a model compound for chlorothiophenols, we found that apart from the mostly discussed chlorothiophenoxy radicals, substituted phenyl radicals and substituted thiophenoxyl diradicals could also be readily formed via the reaction of 2-CTP with H radicals. Furthermore, direct self- and cross-coupling of these radicals can result in the formation of PCDT/TAs, including 1-monochlorothianthrene (1-MCTA), 1,6-dichlorothianthrene (1,6-DCTA), 4,6-dichlorodibenzothiophene (4,6-DCDT) and 1,6-dichlorodibenzothiophene (1,6-DCDT). The pathways proposed in this work are proven to be both thermodynamically and kinetically favorable. Particularly, comparisons were made between the formation mechanisms of sulfurated and oxygenated dioxin systems from an energetic point view, showing that replacing oxygen with sulfur atoms greatly reduces the activation barriers of the rate-controlling steps involved in the PCDT/TA formation processes compared with those involved for PCDD/Fs. The calculated results in this work may improve our understanding of the formation mechanism of PCDT/TAs from chlorothiophenol precursors and should be informative to environmental scientists.