VenomPred: A Machine Learning Based Platform for Molecular Toxicity Predictions

Galati, Salvatore; Stefano, Miriana Di; Martinelli, Elisa; Macchia, Marco; Martinelli, Adriano; Poli, Giulio; Tuccinardi, Tiziano

doi:10.3390/ijms23042105

Cited by 29 publications

(21 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the aim of obtaining robust and reliable toxicity predictions for the four new endpoints, we planned to employ a consensus approach, a strategy that already proved successful in docking and virtual screening studies, 49−52 as well as in MLbased toxicity predictions, as we recently demonstrated. 5 In In the present work, we aimed at maximizing the power of the consensus strategy by performing an exhaustive consensus analysis, evaluating the reliability of as many combinations of ML models as possible to identify the most reliable one. However, we envisioned that employing 20 different models a total of 1,048,555 consensus combinations were possible.…”

Section: ■ Results and Discussionmentioning

confidence: 99%

VenomPred 2.0: A Novel In Silico Platform for an Extended and Human Interpretable Toxicological Profiling of Small Molecules

Di Stefano,

Galati,

Piazza

et al. 2023

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

The application of artificial intelligence and machine learning (ML) methods is becoming increasingly popular in computational toxicology and drug design; it is considered as a promising solution for assessing the safety profile of compounds, particularly in lead optimization and ADMET studies, and to meet the principles of the 3Rs, which calls for the replacement, reduction, and refinement of animal testing. In this context, we herein present the development of VenomPred 2.0 (http://www.mmvsl.it/wp/venompred2/), the new and improved version of our free of charge web tool for toxicological predictions, which now represents a powerful web-based platform for multifaceted and human-interpretable in silico toxicity profiling of chemicals. VenomPred 2.0 presents an extended set of toxicity endpoints (androgenicity, skin irritation, eye irritation, and acute oral toxicity, in addition to the already available carcinogenicity, mutagenicity, hepatotoxicity, and estrogenicity) that can be evaluated through an exhaustive consensus prediction strategy based on multiple ML models. Moreover, we also implemented a new utility based on the Shapley Additive exPlanations (SHAP) method that allows human interpretable toxicological profiling of small molecules, highlighting the features that strongly contribute to the toxicological predictions in order to derive structural toxicophores.

show abstract

Section: ■ Results and Discussionmentioning

confidence: 99%

VenomPred 2.0: A Novel In Silico Platform for an Extended and Human Interpretable Toxicological Profiling of Small Molecules

Di Stefano,

Galati,

Piazza

et al. 2023

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

show abstract

“…7 4; V = 0 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 F = 2; V = 2 The entire scientific community is interested in a quick and accurate method for evalating new compounds' toxicity. Since in vitro and in vivo approaches are frequently conrained by ethics, time, money, and other resources, in silico toxicity prediction methods lay a significant role in the selection of lead drugs and in ADMET research [42]. Conseuently, in this investigation, machine learning (ML) models [43] were used to assess the ossible hepatotoxicity, carcinogenicity, immunotoxicity, mutagenicity, and cytotoxicity f the stearate esters in order to see how these five toxicological elements of drug discovy and development might be affected.…”

Section: Adme/t Analysismentioning

confidence: 99%

Rhamnopyranoside-Based Fatty Acid Esters as Antimicrobials: Synthesis, Spectral Characterization, PASS, Antimicrobial, and Molecular Docking Studies

et al. 2023

View full text Add to dashboard Cite

The most widely used and accessible monosaccharides have a number of stereogenic centers that have been hydroxylated and are challenging to chemically separate. As a result, the task of regioselective derivatization of such structures is particularly difficult. Considering this fact and to get novel rhamnopyranoside-based esters, DMAP-catalyzed di-O-stearoylation of methyl α-l-rhamnopyranoside (3) produced a mixture of 2,3-di-O- (4) and 3,4-di-O-stearates (5) (ratio 2:3) indicating the reactivity of the hydroxylated stereogenic centers of rhamnopyranoside as 3-OH > 4-OH > 2-OH. To get novel biologically active rhamnose esters, di-O-stearates 4 and 5 were converted into six 4-O- and 2-O-esters 6–11, which were fully characterized by FT-IR, 1H, and 13C NMR spectral techniques. In vitro antimicrobial assays revealed that fully esterified rhamnopyranosides 6–11 with maximum lipophilic character showed better antifungal susceptibility than antibacterial activity. These experimental findings are similar to the results found from PASS analysis data. Furthermore, the pentanoyl derivative of 2,3-di-O-stearate (compound 6) showed better antifungal functionality against F. equiseti and A. flavus, which were found to be better than standard antibiotics. To validate the better antifungal results, molecular docking of the rhamnose esters 4–11 was performed with lanosterol 14α-demethylase (PDB ID: 3LD6), including the standard antifungal antibiotics ketoconazole and fluconazole. In this instance, the binding affinities of 10 (−7.6 kcal/mol), 9 (−7.5 kcal/mol), and 7 (−6.9 kcal/mol) were better and comparable to fluconazole (−7.3 kcal/mol), indicating the likelihood of their use as non-azole type antifungal drugs in the future.

show abstract

“…Interestingly, both receptors are associated with melanoma, as well as with many other cancers [114]. In addition, the full toxicology profile of the compounds analyzed, including probabilities of mutagenicity, carcinogenicity, hepatotoxicity, and estrogenicity, was calculated by the VenomPred web software [106]. Probability values are divided into four ranges: <25%, high confidence in the prediction of a molecule as nontoxic, from 25-50%-low confidence in nontoxicity, from 50-75%-low confidence in toxicity, and >75%-high confidence in toxicity.…”

Section: In Silico Pharmacokinetics and Bioactivity Studymentioning

confidence: 99%

“…Additional data were calculated using the pkCSM platform [108]. Full toxicology profiles of analyzed compounds, including probabilities of mutagenicity, carcinogenicity, hepatotoxicity, and estrogenicity were calculated by the VenomPred web tool [106].…”

Section: In Silico Analysismentioning

confidence: 99%

Structural and Biofunctional Insights into the Cyclo(Pro-Pro-Phe-Phe-) Scaffold from Experimental and In Silico Studies: Melanoma and Beyond

Bojarska

Breza

Remko

et al. 2022

IJMS

View full text Add to dashboard Cite

Short peptides have great potential as safe and effective anticancer drug leads. Herein, the influence of short cyclic peptides containing the Pro-Pro-Phe-Phe sequence on patient-derived melanoma cells was investigated. Cyclic peptides such as cyclo(Leu-Ile-Ile-Leu-Val-Pro-Pro-Phe-Phe-), called CLA, and cyclo(Pro-homoPro-β3homoPhe-Phe-), called P11, exert the cytotoxic and the cytostatic effects in melanoma cells, respectively. CLA was the most active peptide as it reduced the viability of melanoma cells to 50% of control at about 10 µM, whereas P11 at about 40 µM after 48 h incubation. Interestingly, a linear derivative of P11 did not induce any effect in melanoma cells confirming previous studies showing that cyclic peptides exert better biological activity compared to their linear counterparts. According to in silico predictions, cyclic tetrapeptides show a better pharmacokinetic and toxic profile to humans than CLA. Notably, the spatial structure of those peptides containing synthetic amino acids has not been explored yet. In the Cambridge Structural Database, there is only one such cyclic tetrapeptide, cyclo((R)-β2homoPhe-D-Pro-Lys-Phe-), while in the Protein Data Bank—none. Therefore, we report the first crystal structure of cyclo(Pro-Pro-β3homoPhe-Phe-), denoted as 4B8M, a close analog of P11, which is crucial for drug discovery. Comparative molecular and supramolecular analysis of both structures was performed. The DFT findings revealed that 4B8M is well interpreted in the water solution. The results of complex Hirshfeld surface investigations on the cooperativity of interatomic contacts in terms of electrostatic and energetic features are provided. In short, the enrichment ratio revealed O…H/H…O and C…H/H…C as privileged intercontacts in the crystals in relation to basic and large supramolecular H-bonding synthon patterns. Furthermore, the ability of self-assemble 4B8M leading to a nanotubular structure is also discussed.

show abstract

VenomPred: A Machine Learning Based Platform for Molecular Toxicity Predictions

Cited by 29 publications

References 25 publications

VenomPred 2.0: A Novel In Silico Platform for an Extended and Human Interpretable Toxicological Profiling of Small Molecules

VenomPred 2.0: A Novel In Silico Platform for an Extended and Human Interpretable Toxicological Profiling of Small Molecules

Rhamnopyranoside-Based Fatty Acid Esters as Antimicrobials: Synthesis, Spectral Characterization, PASS, Antimicrobial, and Molecular Docking Studies

Structural and Biofunctional Insights into the Cyclo(Pro-Pro-Phe-Phe-) Scaffold from Experimental and In Silico Studies: Melanoma and Beyond

Contact Info

Product

Resources

About