Holistic Prediction of the pK_a in Diverse Solvents Based on a Machine‐Learning Approach

Abstract

While many approaches to predict aqueous pK a values exist, the fast and accurate prediction of non-aqueous pK a values is still challenging. Based on the iBonD experimental pK a database (39 solvents), ah olistic pK a prediction model was established using machine learning.S tructural and physical-organic-parameter-based descriptors (SPOC) were introduced to represent the electronic and structural features of the molecules.The models trained with aneural network or the XGBoost algorithm showed the best prediction performance with alow MAE value of 0.87 pK a units.The approachallows ac omprehensive mapping of all possible pK a correlations between different solvents and it was validated by predicting the aqueous pK a and micro-pK a of pharmaceutical molecules and pK a values of organocatalysts in DMSO and MeCN with high accuracy.Anonline prediction platform was constructed based on the current model, which can providepK a prediction for different types of XÀHacidity in the most commonly used solvents.

MolGpka: A Web Server for Small Molecule pKa Prediction Using a Graph-Convolutional Neural Network

“…5,6,14 The LFER models apply the Hammett equations to predict pK a by classifying the molecule to a parent class and modifying the pK a value of the parent class with a property of its substituents. Machine learning models 11,12,15 usually use a molecular environment rooted at the ionization center as the descriptor to develop the pK a prediction approach by learning from data.…”

Thermodynamic and Kinetic Studies of the Activities of Aldehydic C−H Bonds toward Their H‐Atom Transfer Reactions

“…According to the definition of Δ G ≠ °(XH), as long as the bond dissociation free energies Δ G o (XH) of XH are available, the intrinsic resistance energies Δ G ≠ XH/X can be obtained by eq 3. In this work, the bond dissociation energies (BDE) of C(sp 2 )−H bonds of these eight aldehydes in acetonitrile were calculated using the i BonD HM method developed by Luo and Zhang in 2020, whose details are available at http://pka.luoszgroup.com/bde_prediction [23,24] . It's a holistic BDE prediction model (HM) based on the i BonD experimental dataset [25] .…”

“…Thermodynamic Measurements . The bond dissociation free energies Δ G o (XH) are determined by the i BonD HM prediction method: Light GBM and SPOC descroptors with RMSE=1.82, r 2 =0.980 and Mean Absolute Errors (MAE)=1.03 (95 : 5 train test split) [23,24] …”

Crystallographic Deciphering of Spontaneous Self‐Assembly of Achiral Calciphores to Chiral Complexes

“…The analysis of the one‐dimensional and 2D NMR spectra indicated the absence of one methyl group in the aliphatic section compared with that of 1 (Figure 1d). To confirm the aforementioned structures, a positive‐model ESI‐MS/MS experiment was performed, in which the machine‐learning predictions of the bond dissociation energy were introduced [29] . Notably, the vital details of the cleavage pattern were established by the careful examination of the (+)‐ESI‐MS/MS spectra, in which the diagnostic peaks at m/z 236.0247 and 210.0217 indicate the common moiety shared by the two different compounds; the results were consistent with those of the prediction data (Tables S2–S3, Figures.…”

“…Based on Chen's machine‐learning calculation approach, [29] the acidity of the TPCs was predicted as pK a 7.59 and 7.72 for 1 and 2 (Figure S26, Table S8), respectively; the pK a value indicated that the TPCs have a low tendency to dissociate into monoanionic TPCs by losing the protons from the C5‐OH. A lack of charged solvent molecules to neutralize the charge field in the crystal packing was observed; therefore, a coordination scaffold was proposed.…”