Using Background Knowledge from Preceding Studies for Building a Random Forest Prediction Model: A Plasmode Simulation Study

There is an increasing interest in machine learning (ML) algorithms for predicting patient outcomes, as these methods are designed to automatically discover complex data patterns. For example, the random forest (RF) algorithm is designed to identify relevant predictor variables out of a large set of candidates. In addition, researchers may also use external information for variable selection to improve model interpretability and variable selection accuracy, thereby prediction quality. However, it is unclear to which extent, if at all, RF and ML methods may benefit from external information. In this paper, we examine the usefulness of external information from prior variable selection studies that used traditional statistical modeling approaches such as the Lasso, or suboptimal methods such as univariate selection. We conducted a plasmode simulation study based on subsampling a data set from a pharmacoepidemiologic study with nearly 200,000 individuals, two binary outcomes and 1152 candidate predictor (mainly sparse binary) variables. When the scope of candidate predictors was reduced based on external knowledge RF models achieved better calibration, that is, better agreement of predictions and observed outcome rates. However, prediction quality measured by cross-entropy, AUROC or the Brier score did not improve. We recommend appraising the methodological quality of studies that serve as an external information source for future prediction model development.     

Reactivity of alkynes containing alpha-hydrogen atoms with a triruthenium hydrido carbonyl cluster: alkenyl versus allyl cluster derivatives

The reactions of the hydrido-triruthenium cluster complex [Ru3(mu-H)(mu3-kappa(2)-HNNMe2)(CO)9] (1; H2NNMe2 = 1,1-dimethylhydrazine) with alkynes that have alpha-hydrogen atoms give trinuclear derivatives containing edge-bridging allyl or face-capping alkenyl ligands. Under mild conditions (THF, 70 degrees C) the isolated products are as follows: [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(3)-1-syn-Me-3-anti-EtC3H3)(mu-CO)2(CO)6] (2) and [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(3)-1-syn-Me-3-syn-EtC3H3)(mu-CO)2(CO)6] (3) from 3-hexyne; [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(3)-3-anti-PhC3H4)(mu-CO)2(CO)6] (4), [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(2)-MeCCHPh)(mu-CO)2(CO)6] (5) and [Ru3(mu3-kappa(2)-HNNMe2)(mu3-kappa(2)-PhCCHMe)(mu-CO)2(CO)6] (6) from 1-phenyl-1-propyne; [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(2)-3-anti-PrC3H4)(mu-CO)2(CO)6] (7), [Ru3(mu3-kappa(2)-HNNMe2)(mu3-kappa(2)-BuCCH2)(mu-CO)2(CO)6] (8), and [Ru3(mu3-kappa(2)-HNNMe2)(mu3-kappa(2)-HCCHBu)(mu-CO)2(CO)6] (9) from 1-hexyne; [Ru3(mu3-kappa(2)-HNNMe2)(mu3-kappa(2)-HOH2CCCH2)(mu-CO)2(CO)6] (10) from propargyl alcohol; and [Ru3(mu3-kappa(2)-HNNMe2)(mu3-kappa(2)-MeOCH2CCH2)(mu-CO)2(CO)6] (11) from 3-methoxy-1-propyne. The regioselectivity of these reactions depends upon the nature of the alkyne reagent, which affects considerably the kinetic barriers of important reaction steps and the stability of the final products. It has been established that the face-capped alkenyl derivatives are not precursors to the allyl products, which are formed via edge-bridged alkenyl intermediates. At higher temperature (toluene, 110 degrees C), the complexes that have allyl ligands with an anti substituent are isomerized into allyl derivatives with that substituent in the syn position, for example, 4 into [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(3)-3-syn-PhC3H4)(mu-CO)2(CO)6] (14). The diene complex [Ru3(mu-H)(mu3-kappa(2)-HNNMe2)(mu-kappa(4)-trans-EtC4H5)(CO)7] (13) has been obtained from the thermolysis of compounds 2 and 7 at 110 degrees C (3 and [Ru3(mu3-kappa(2)-HNNMe2)(mu-kappa(2)-3-syn-PrC3H4)(mu-CO)2(CO)6] (12) are also formed in these reactions). A DFT theoretical study has allowed a comparison of the thermodynamic stabilities of isomeric compounds and has helped rationalize the experimental results. Mechanistic proposals for the synthesis of the allyl complexes and their isomerization processes are also provided.     

Solid-phase reducing agents as alternative for reducing disulfide bonds in proteins

Disulfide reduction of Kluyveromyces lactis and Aspergillus oryzae β-galactosidases and β-lactoglobulin was assessed. Reduction was performed using one of two thiol-containing agents: dithiothreitol (DTT) or thiopropyl-agarose with a high degree of substitution (1000 μmol of SH groups/g of dried gel). Both reductants allowed an increase of three- (for K. lactis β-galactosidase) and fourfold (for A. oryzae β-galactosidase) in the initial content of SH groups in the lactases. Nearly sevenfold fewer micromoles of SH groups per milligram of protein were needed to perform the reduction of K. lactis β-galactosidase with thiopropyl-agarose than for the same reduction with DTT. However, for A. oryzae β-galactosidase, nearly twice as many micromoles of SH groups per milligram of protein were needed with thiopropylagarose than with DTT. Disulfide bonds in β-lactoglobulin were not accessible to thiopropyl-agarose, since this reduction was only possible in the presence of 6 M urea. These results proved that highly substituted thiopropyl-agarose is as good a reducing agent as DTT, for the reduction of disulfide bonds in proteins. Moreover, excess reducing agent was very simply separated from the reduced protein by filtration, making it easier to control the reaction and providing reduced protein solutions free of reductant. All these advantages substantially cut down the time required and therefore the cost of the overall process.     

Towards a Precise NMR Quantification of Acute Phase Inflammation Proteins from Human Serum

Nuclear Magnetic Resonance (NMR) spectra of human serum and plasma show, besides metabolites and lipoproteins, two characteristic signals termed GlycA and B arising from the acetyl groups of glycoprotein glycans from acute phase proteins, which constitute good markers for inflammatory processes. Here, we report a comprehensive assignment of glycoprotein glycan NMR signals observed in human serum, showing that GlycA and GlycB signals originate from Neu5Ac and GlcNAc moieties from N-glycans, respectively. Diffusion-edited NMR experiments demonstrate that signal components can be associated with specific acute phase proteins. Conventionally determined concentrations of acute phase glycoproteins correlate well with distinct features in NMR spectra (R² up to 0.9422, p-value <0.001), allowing the simultaneous quantification of several acute phase inflammation proteins. Overall, a proteo-metabolomics NMR signature of significant diagnostic potential is obtained within 10–20 min acquisition time. This is exemplified in serum samples from COVID-19 and cardiogenic shock patients showing significant changes in several acute phase proteins compared to healthy controls.     

Identification of cellular pathways affected by Sortin2, a synthetic compound that affects protein targeting to the vacuole in Saccharomyces cerevisiae

Background

Sortin2 is a low mass compound that interferes with vacuolar delivery of proteins in plants and yeast. The Sortin2 phenotype was tested in Arabidopsis thaliana and found to be reversible upon drug removal, demonstrating the ability of chemical genomics to induce reversible phenotypes that would be difficult to achieve using conventional genetics [1]. However, standard genetic methods can be used to identify drug target pathways in a high-throughput manner.

Results

In this study, we analyzed structure-function relationships of Sortin2 using structural analogues. The results show the key roles of sulphite substitution and a benzoic acid group. A Sortin 2 hypersensitivity screen for the induced secretion of a vacuolar cargo protein was done utilizing a yeast haploid deletion library. Using bioinformatics approaches, we highlighted functional information about the cellular pathways affected by drug treatment which included protein sorting and other endomembrane system-related processes.

Conclusion

Chemical, genomic and genetics approaches were used to understand the mode of action of Sortin2, a bioactive chemical that affects the delivery of a vacuolar protein. Critical features of Sortin2 structure necessary for bioactivity suggest a binding pocket that may recognize two ends of Sortin2. The genome-wide screen shows that Sortin2 treatment in yeast affects primarily components within the endomembrane system. This approach allowed us to assign putative functions in protein sorting for fifteen genes of previously unknown function.     

Could Light-Based Technologies Improve Stem Cell Therapy for Skin Wounds? A Systematic Review and Meta-Analysis of Preclinical Studies†

Several diseases or conditions cause dermatological disorders that hinder the process of skin repair. The search for novel technologies has inspired the combination of stem cell (SC) and light-based therapies to ameliorate skin wound repair. Herein, we systematically revised the impact of photobiomodulation therapy (PBM) combined with SCs in animal models of skin wounds and quantitatively evaluated this effect through a meta-analysis. For inclusion, SCs should be irradiated in vitro or in vivo, before or after being implanted in animals, respectively. The search resulted in nine eligible articles, which were assessed for risk of bias. For the meta-analysis, studies were included only when PBM was applied in vivo, five regarding wound closure, and three to wound strength. Overall, a positive influence of SC + PBM on wound closure (mean difference: 9.69; 95% CI: 5.78–13.61, P < 0.00001) and strength (standardized mean difference: 1.7, 95% CI: 0.68–2.72, P = 0.001) was detected, although studies have shown moderate to high heterogeneity and a lack of information regarding some bias domains. Altogether, PBM seems to be an enabling technology able to be applied postimplantation of SCs for cutaneous regeneration. Our findings may guide future laboratory and clinical studies in hopes of offering wound care patients a better quality of life.     

MoBioTools: A toolkit to setup quantum mechanics/molecular mechanics calculations

We present a toolkit that allows for the preparation of QM/MM input files from a conformational ensemble of molecular geometries. The package is currently compatible with trajectory and topology files in Amber, CHARMM, GROMACS and NAMD formats, and has the possibility to generate QM/MM input files for Gaussian (09 and 16), Orca (≥4.0), NWChem and (Open)Molcas. The toolkit can be used in command line, so that no programming experience is required, although it presents some features that can also be employed as a python application programming interface. We apply the toolkit in four situations in which different electronic-structure properties of organic molecules in the presence of a solvent or a complex biological environment are computed: the reduction potential of the nucleobases in acetonitrile, an energy decomposition analysis of tyrosine interacting with water, the absorption spectrum of an azobenzene derivative integrated into a voltage-gated ion channel, and the absorption and emission spectra of the luciferine/luciferase complex. These examples show that the toolkit can be employed in a manifold of situations for both the electronic ground state and electronically excited states. It also allows for the automatic correction of the active space in the case of CASSCF calculations on an ensemble of geometries, as it is shown for the azobenzene derivative photoswitch case.     

Synthesis of Weinreb amides via Pd-catalyzed aminocarbonylation of heterocyclic-derived triflates

The direct transformation of lactam-, lactone-, and thiolactone-derived triflates into N-methoxy-N-methyl or morpholine Weinreb amides has been realized using Pd-catalyzed aminocarbonylation under CO atmospheric pressure and at room temperature. The carbonylative coupling can be efficiently carried out with 2% of catalyst in the presence of Xantphos as a ligand. The amides smoothly react with nucleophiles to afford acylated aza-, oxa-, and thio-heterocycles. The proposed methodology could be advantageously exploited for the synthesis of dienones in which one of the double bonds is embedded in a heterocyclic moiety, as useful substrates for Nazarov cyclization.     

Chemical and biological profiles of novel copper(II) complexes containing S-donor ligands for the treatment of cancer

In the last years, we have synthesized some new platinum(II), palladium(II), gold(I/III) complexes with dithiocarbamato derivatives as potential anticancer drugs, to obtain compounds with superior chemotherapeutic index in terms of increased bioavailability, higher cytotoxicity, and lower side effects than cisplatin. On the basis of the obtained encouraging results, we have been studying the interaction of CuCl2 with methyl-/ethyl-/tert-butylsarcosine-dithiocarbamato moieties in a 1:2 molar ratio; we also synthesized and studied the N,N-dimethyl- and pyrrolidine-dithiocarbamato copper complexes for comparison purposes. The reported compounds have been successfully isolated, purified, and fully characterized by means of several spectroscopic techniques. Moreover, the electrochemical properties of the designed compounds have been studied through cyclic voltammetry. In addition, the behavior in solution was followed by means of UV-vis technique to check the stability with time in physiological conditions. To evaluate their in vitro cytotoxic properties, preliminary biological assays (MTT test) have been carried out on a panel of human tumor cell lines. The results show that cytotoxicity levels of all of the tested complexes are comparable or even greater than that of the reference drug (cisplatin).     

Interactions of aminomethylphosphonic acid and sarcosine with montmorillonite interlayer surfaces

The smectite clay, montmorillonite, can be found in many soils throughout the world. In addition to its importance in agriculture and soil remediation, montmorillonite has extensive applications in industry both in its natural form and as a component of composite materials. The adsorptive properties of montmorillonite have been explored in relation to its interactions with the common herbicide glyphosate. This herbicide, when exposed to microorganisms in the soil is degraded, forming two products: aminomethylphosphonic acid (AMPA) and sarcosine. The atomic‐level interactions of these compounds with the montmorillonite interlayer surfaces are studied here using molecular mechanics. The final outcomes of these calculations are analyzed in terms of the proximity of the montmorillonite surface to the moieties of the degradation products. The phosphonate moiety was found to be the most important source of interactions for AMPA, while for sarcosine there was an even distribution between the amino and carboxylic moieties, and Na⁺ ion mediated surface complexes. © 2007 Wiley Periodicals, Inc. Int J Quantum Chem, 2008