Application of linear solvation energy relationships and principal component analysis methods for the prediction of the retention behaviour of E-resveratrol analogues with substituted silica hydride stationary phases

Evaluation of the permeability and potential toxicity of polycyclic aromatic hydrocarbons to pulmonary surfactant membrane by the parallel artificial membrane permeability assay model

Polycyclic aromatic hydrocarbons (PAHs) can penetrate and accumulate in the pulmonary surfactant (PS) membranes, leading to abnormalities of biological macromolecules and the destruction of membrane structure and properties. In the present study, the bioavailability, apparent permeability, effective permeability and residual coefficient of 10 PAHs on PS membrane was assessed by the parallel artificial membrane permeability assay (PAMPA). The influence of various forces on permeability is obtained by analyzing the correlation between parameters and physicochemical properties. Research shows that octanol-water partition coefficient (K_ow) cannot directly predict permeability, and permeability has no significant relationship with polarity. Dispersion, induction, coupling/polarization promote permeation, while hydrogen bonded acid and n-n electron pair inhibit permeation. Further surface pressure-area (π-A) isotherms test and Brewster angle microscope observation manifested that there are huge differences in the transmembrane ability and effects on the membrane of PAHs with different structures. This work has considerable significance that will help to evaluate the bioavailability and human health risk of PAHs.

Retention time prediction in hydrophilic interaction liquid chromatography with graph neural network and transfer learning

The combination of retention time (RT), accurate mass and tandem mass spectra can improve the structural annotation in untargeted metabolomics. However, the incorporation of RT for metabolite identification has received less attention because of the limitation of available RT data, especially for hydrophilic interaction liquid chromatography (HILIC). Here, the Graph Neural Network-based Transfer Learning (GNN-TL) is proposed to train a model for HILIC RTs prediction. The graph neural network was pre-trained using an in silico HILIC RT dataset (pseudo-labeling dataset) with ∼306 K molecules. Then, the weights of dense layers in the pre-trained GNN (pre-GNN) model were fine-tuned by transfer learning using a small number of experimental HILIC RTs from the target chromatographic system. The GNN-TL outperformed the methods in Retip, including the Random Forest (RF), Bayesian-regularized neural network (BRNN), XGBoost, light gradient-boosting machine (LightGBM), and Keras. It achieved the lowest mean absolute error (MAE) of 38.6 s on the test set and 33.4 s on an additional test set. It has the best ability to generalize with a small performance difference between training, test, and additional test sets. Furthermore, the predicted RTs can filter out nearly 60% false positive candidates on average, which is valuable for the identification of compounds complementary to mass spectrometry.

Synthesis of octadecylamine-derived carbon dots and application in reversed phase/hydrophilic interaction liquid chromatography

In order to make up for the deficiencies of traditional C18 column for separating strong polar compounds, combined with the good hydrophilicity of carbon dots (CDs), novel octadecylamine-derived CDs denoted as C18-CDs are designed, synthesized and applied in RPLC/HILIC mixed-mode chromatography with good separation performance towards both hydrophobic and hydrophilic compounds. C18-CDs are synthesized by simple one-step solvothermal method using octadecylamine and citric acid as carbon sources, and C18-CDs with proper polarity are collected through column chromatography purification. This C18-CDs decorated silica column showed good separation performance for polycyclic aromatic hydrocarbons and alkylbenzenes under RPLC mode. Hydrophilic compounds including sulfonamides, nucleosides and nucleobases also achieved good resolution in HILIC mode. Hydrophobic and π-π stacking interactions play major retaining roles in RPLC, whereas hydrophilic partitioning and hydrogen bond interactions turn to the main retention interactions under HILIC mode. This C18-CDs/SiO₂ column was applied for the fast detection of chloramphenicol in milk without complex sample pretreatment process. Quantitative relationship between the peak area and the concentration of chloramphenicol was established with linear equation of A = 1677c + 173. Satisfactory spiked recoveries in the range of 94.1-109.0% were obtained. This work not only proposes a simple method for improving the polarity of C18 column through forming octadecane into CDs, but also provides novel CDs with certain hydrophobicity/hydrophily suitable for mixed-mode chromatography.