Supervised machine learning algorithms for protein structure classification

We explore automation of protein structural classification using supervised machine learning methods on a set of 11,360 pairs of protein domains (up to 35% sequence identity) consisting of three secondary structure elements. Fifteen algorithms from five categories of supervised algorithms are evaluated for their ability to learn for a pair of protein domains, the deepest common structural level within the SCOP hierarchy, given a one-dimensional representation of the domain structures. This representation encapsulates evolutionary information in terms of sequence identity and structural information characterising the secondary structure elements and lengths of the respective domains. The evaluation is performed in two steps, first selecting the best performing base learners and subsequently evaluating boosted and bagged meta learners. The boosted random forest, a collection of decision trees, is found to be the most accurate, with a cross-validated accuracy of 97.0% and F-measures of 0.97, 0.85, 0.93 and 0.98 for classification of proteins to the Class, Fold, Super-Family and Family levels in the SCOP hierarchy. The meta learning regime, especially boosting, improved performance by more accurately classifying the instances from less populated classes.     

Implementation of machine learning algorithms for detecting missing radioactive material

Journal of Radioanalytical and Nuclear Chemistry - The detection of missing radioactive material is an important capability for safeguards measurements. Gamma ray signatures provide sample...     

Identification of coenzyme-binding proteins with machine learning algorithms

The coenzyme-binding proteins play a vital role in the cellular metabolism processes, such as fatty acid biosynthesis, enzyme and gene regulation, lipid synthesis, particular vesicular traffic, and β-oxidation donation of acyl-CoA esters. Based on the theory of Star Graph Topological Indices (SGTIs) of protein primary sequences, we proposed a method to develop a first classification model for predicting protein with coenzyme-binding properties. To simulate the properties of coenzyme-binding proteins, we created a dataset containing 2897 proteins, among 456 proteins functioned as coenzyme-binding activity. The SGTIs of peptide sequence were calculated with Sequence to Star Network (S2SNet) application. We used the SGTIs as inputs to several classification techniques with a machine learning software - Weka. A Random Forest classifier based on 3 features of the embedded and non-embedded graphs was identified as the best predictive model for coenzyme-binding proteins. This model developed was with the true positive (TP) rate of 91.7%, false positive (FP) rate of 7.6%, and Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.971. The prediction of new coenzyme-binding activity proteins using this model could be useful for further drug development or enzyme metabolism researches.     

Surface-bonded ionic liquid stationary phases in high-performance liquid chromatography—A review

Ionic liquids (ILs) are a class of ionic, nonmolecular solvents which remain in liquid state at temperatures below 100 °C. ILs possess a variety of properties including low to negligible vapor pressure, high thermal stability, miscibility with water or a variety of organic solvents, and variable viscosity. IL-modified silica as novel high-performance liquid chromatography (HPLC) stationary phases have attracted considerable attention for their differential behavior and low free-silanol activity. Indeed, around 21 surface-confined ionic liquids (SCIL) stationary phases have been developed in the last six years. Their chromatographic behavior has been studied, and, despite the presence of a positive charge on the stationary phase, they showed considerable promise for the separation of neutral solutes (not only basic analytes), when operated in reversed phase mode. This aspect points to the potential for truly multimodal stationary phases. This review attempts to summarize the state-of-the-art about SCIL phases including their preparation, chromatographic behavior, and analytical performance.     

Na-K liquid alloy:A review on wettability enhancement and ionic carrier selection mechanism

The intrinsic liquid interface of Na-K alloy allays concerns about dendrite growth on metal anodes that are thermodynamically within the room temperature(20-22℃).Nevertheless,it hinders the formation of a stable electrode structure due to the inferior wettability induced by considerable liquid tension.In addition,the dominant ionic carrier in the Na-K alloy is subject to multiple factors,which is not conducive to customized battery design.This review,based on recently reported frontier achievements on Na-K liquid anodes,summarizes practical strategies for promoting the wettability by hightemperature induction,capillary effect,vacuum infiltration,and solid interface protection.Furthermo re,four selection mechanisms of the dominant ionic carrier are presented:(1) ion property dominated,(2)cathode dominated,(3) separator dominated,and(4) solid electrolyte interface dominated.Notably,initial electrolytes in energy storage systems have been unable to play a decisive role in ionic selection.Utilizing a superior wettability strategy and simultaneously identifying the dominant ionic carrier can facilitate the tailored application of dendrite-free Na-K liquid anodes.     

Influence of the ionic liquid/gas surface on ionic liquid chemistry

Applications such as gas storage, gas separation, NP synthesis and supported ionic liquid phase catalysis depend upon the interaction of different species with the ionic liquid/gas surface. Consequently, these applications cannot proceed to the full extent of their potential without a profound understanding of the surface structure and properties. As a whole, this perspective contains more questions than answers, which demonstrates the current state of the field. Throughout this perspective, crucial questions are posed and a roadmap is proposed to answer these questions. A critical analysis is made of the field of ionic liquid/gas surface structure and properties, and a number of design rules are mined. The effects of ionic additives on the ionic liquid/gas surface structure are presented. A possible driving force for surface formation is discussed that has, to the best of my knowledge, not been postulated in the literature to date. This driving force suggests that for systems composed solely of ions, the rules for surface formation of dilute electrolytes do not apply. The interaction of neutral additives with the ionic liquid/gas surface is discussed. Particular attention is focussed upon H(2)O and CO(2), vital additives for many applications of ionic liquids. Correlations between ionic liquid/gas surface structure and properties, ionic liquid surfaces plus additives, and ionic liquid applications are given.     

A greener electrochromic liquid crystal based on ionic liquid electrolytes

An electrochromic liquid crystal (ECLC) material composed of only liquid crystal (LC) and ionic liquid (IL) was developed. The LC containing the substituted diphenylacetylene serves as electrochromic (EC) material to realise transmittance and colour change under the direct current (DC) field, while the IL with the designable cation and anion served as electrolyte. Herein, a series of IL electrolytes was screened to investigate how IL tunes the electro-optic performance of the ECLC cell. By testing the electrochemistry window of ILs in EC cells, IL with the [NTf₂]^? anion shows adequate electrochemical stability when the EC material undergoes oxidation and reduction. The electro-optic performance of ECLC containing 1-ethoxy-4-[2-(4-pentylphenyl) ethynyl]-benzene (PEB) and IL was then evaluated by UV-vis spectrometry under the control of an electrochemical work station. Compared with other PEB-IL, PEB-[Bmim][NTf₂] with [Bmim][NTf₂] electrolyte shows a satisfactory transmittance at low operating voltage. Furthermore, Pd NPs in situ formed in [Bmim][NTf₂] reduced the EC potential and improved the light scattering of the ECLC cell. In this work, we also designed a bifunctional device based on polymer dispersed liquid crystal (PDLC) that hosts electrochromic guest molecules, and analysed the electro-optical and electrochromic properties of LC electrolyte mixtures, in order to gain control of the incident daylight and glare in building and automotive applications.     

A linear learning machine program for a microcomputer

The linear learning machine (LLM) has become an established procedure in chemical pattern recognition. It examines whether two classes of objects, represented as points in a multidimensional space with the variables as coordinates, can be separated by means of a linear hyperplane. The position of the hyperplane is determine by an iterative procedure, using feedback corrections. In most applications, the computations are done on minicomputers. In this communication, a “micro-LLM” is described, i.e., a linear learning machine implemented on a microcomputer.     

A cross-linker containing aldehyde functionalized ionic liquid for chitosan

The synthesis of conductive and eco-friendly new chitosan composite biopolymer with a novel aldehyde functionalized ionic liquids (AFILs) was investigated in this work. The AFILs, 1,1′-(pentane-1,5-diyl) bis(4-formylpyridin-1-ium) dibromide (C₅PyBr), 1,1′-(octane-1,8-diyl) bis(4-formylpyridin-1-ium) dibromide (C₈PyBr) and 1,1′-(decane-1,10-diyl) bis(4-formylpyridin-1-ium) dibromide (C₁₀PyBr) have been synthesized and they were characterized by FT-IR, UV–vis and ¹H-NMR and ¹³ C-NMR, TG, and DSC analyses. The synthesized AFILs, C₅PyBr C₈PyBr and C₁₀PyBr were used as cross-linker for chitosan biopolymers and the cross-linked chitosan composites, chitosan/1,1′-(pentane-1,5-diyl)bis(4-formylpyridin-1-ium)dibromide (Ch/C₅PyBr), chitosan/1,1′-(octane-1,8-diyl)bis(4-formylpyridin-1-ium)dibromide (Ch/C₈PyBr) and chitosan/1,1′-(decane-1,10-diyl)bis(4-formylpyridin-1-ium)dibromide (Ch/C₁₀PyBr) were prepared. Herein, the synthesized AFILs have not only cross-linked chitosan, but also provide them the desired novel, functional groups. The prepared chitosan composites were characterized by FT-IR for the analysis of structural, TG and DSC for the thermal behavior, and an electrometer for the electrical conductivity. The electrical conductivity of the prepared Ch/C₅PyBr composite was measured as 1.69?×?10^?5 ± 2.37?×?10^?5 S cm^?1 by the four-point probe at room temperature and this value is higher about 4500-fold than the electrical conductivity of the chitosan.     

A ionic liquid for dispersive liquid-liquid microextraction of phenols

In the present study, a new solvent-free mode of liquid phase microextraction termed ionic liquid dispersive liquid-liquid microextraction (IL-DLLME) was developed. Four phenols were used as model compounds in the development and evaluation of the procedure. In this method, 50 μL of ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate ([Bmim][PF₆]) and 1.5 mL of sample aqueous solution were placed in a 2.2-mL glass test tube and mixed by aspirating and rapidly injecting by a syringe. This procedure produced a cloudy solution. In this process, phenols in the water sample were extracted into the IL phase. After centrifuging, the fine droplets of IL sedimented to the bottom of the glass test tube. The settled phase was injected into the high performance liquid chromatograph (HPLC) for separation and detection of phenols. Some parameters that might affect the extraction efficiency were optimized. The main advantages of the proposed method are high speed, high recovery, good repeatability and environmental friendliness.     

A highly active ionic liquid catalyst for Morita-Baylis-Hillman reaction

The Morita-Baylis-Hillman reaction is an efficient carbon-carbon bond forming reaction for the preparation of α-methylene-β-hydroxycarbonyl compounds. A new and highly active di-naphthalene imidazolium salt has been synthesized. We have found that 1,3-bis[2-(naphthalene-2-yloxy)propyl]imidazolium bromid promoted the Morita-Baylis-Hillman reaction of various aryl aldehyde compounds in the absence of solvents. Our studies show that the Morita-Baylis-Hillman reaction by the influence of ionic liquid to give a high yield and short reaction time.     

A green and novel procedure for the preparation of ionic liquid

A green and novel procedure is described for the preparation of a series of ionic liquid containing alkylimidazolium-based or N-alkylpyridinium-based cations and hexafluorophosphate-based or tetrafluoroborate-based anions in one-pot solvent-free conditions to give excellent yields with shortened time.     

Novel algorithms for searching conformational space

Summary The modelling of biological molecules by molecular dynamics is beset by a range of problems. The most important of these is the multiple-minima problem. The deep metastable minima can cause difficulties in proper equilibration of a molecular system and result in the simulated system being trapped in a long-lived metastable state. One way to overcome these problems is to re-engineer the Newtonian Rules in order to more efficiently search conformational space. Re-engineering of the Newtonian Rules implies a redesign of the physical laws arising from them. This is done in various ways by the RUSH, Hybrid Monte Carlo and PEACS algorithms. This paper explores applications of these algorithms, and compares them to a traditional molecular dynamics method.     

Improving machine learning performance on small chemical reaction data with unsupervised contrastive pretraining

Machine learning (ML) methods have great potential to transform chemical discovery by accelerating the exploration of chemical space and drawing scientific insights from data. However, modern chemical reaction ML models, such as those based on graph neural networks (GNNs), must be trained on a large amount of labelled data in order to avoid overfitting the data and thus possessing low accuracy and transferability. In this work, we propose a strategy to leverage unlabelled data to learn accurate ML models for small labelled chemical reaction data. We focus on an old and prominent problem—classifying reactions into distinct families—and build a GNN model for this task. We first pretrain the model on unlabelled reaction data using unsupervised contrastive learning and then fine-tune it on a small number of labelled reactions. The contrastive pretraining learns by making the representations of two augmented versions of a reaction similar to each other but distinct from other reactions. We propose chemically consistent reaction augmentation methods that protect the reaction center and find they are the key for the model to extract relevant information from unlabelled data to aid the reaction classification task. The transfer learned model outperforms a supervised model trained from scratch by a large margin. Further, it consistently performs better than models based on traditional rule-driven reaction fingerprints, which have long been the default choice for small datasets, as well as those based on reaction fingerprints derived from masked language modelling. In addition to reaction classification, the effectiveness of the strategy is tested on regression datasets; the learned GNN-based reaction fingerprints can also be used to navigate the chemical reaction space, which we demonstrate by querying for similar reactions. The strategy can be readily applied to other predictive reaction problems to uncover the power of unlabelled data for learning better models with a limited supply of labels.

Contrastive pretraining of chemical reactions by matching augmented reaction representations to improve machine learning performance on small reaction datasets.     

A novel ionic liquid based on small size cation

An ionic liquid material containing small size cation was synthesized and its composition and structure was determined by FT-IR, 1H NMR, 13C NMR and element analysis. However, its physical properties such as melting point, viscosity, conductivity, etc. were measured and the results show that the novel ionic liquid is composed of small cation with highly symmetry and larger anion which is very rare among the reported ionic liquid family.     

Modeling of liquid fuel purification by the LTA zeolite using machine learning methods

Journal of Thermal Analysis and Calorimetry - Impurities have severe negative influences on both the application and energy content of the dimethyl aminoethyl azide (DMAZ) fuel. Therefore, it is...