Physicochemical Descriptors Research Articles

Premexotac: Machine learning bitterants predictor for advancing pharmaceutical development

Bitter taste receptors were recently found to be involved in numerous physiological and pathological conditions other than taste and are suggested as potential drug targets. In vivo and in vitro techniques for screening bitterants as ligands come with economical, time and ethic challenges. Therefore, in silico tools can represent a valuable alternative due to their practicality. Yet, the main challenge of already established ligand-based (LB) classifiers is the low number of experimentally confirmed bitterants and non-bitterants. Premexotac models were constructed as a LB bitterants screener, exploring novel combinations of feature extraction, feature selection and learning algorithms as a contrast with the already available screeners. Premexotac came among the top performers, exhibiting a F-1 score up to 81% on external validation. Premexotac identified as well insights on physicochemical and topological descriptors important for bitter prediction. Among the key insights, important molecular substructures from Extended Connectivity Fingerprints for bitterness classification were identified. Also, the importance of a selection of physicochemical/topological descriptors was ranked using mutual information and it was found that descriptors related to the ramification of the molecular structure and molecular weight came at the top of the ranking. The remaining challenges for improving performance were discussed and stated, widening the LB bitterness prediction outlook.

A Computational QSAR, Molecular Docking and In Vitro Cytotoxicity Study of Novel Thiouracil-Based Drugs with Anticancer Activity against Human-DNA Topoisomerase II.

Computational chemistry, molecular docking, and drug design approaches, combined with the biochemical evaluation of the antitumor activity of selected derivatives of the thiouracil-based dihydroindeno pyrido pyrimidines against topoisomerase I and II. The IC50 of other cell lines including the normal human lung cell line W138, lung cancer cell line, A549, breast cancer cell line, MCF-7, cervical cancer, HeLa, and liver cancer cell line HepG2 was evaluated using biochemical methods. The global reactivity descriptors and physicochemical parameters were computed, showing good agreement with the Lipinski and Veber’s rules of the drug criteria. The molecular docking study of the ligands with the topoisomerase protein provides the binding sites, binding energies, and deactivation constant for the inhibition pocket. Various biochemical methods were used to evaluate the IC50 of the cell lines. The QSAR model was developed for colorectal cell line HCT as a case study. Four QSAR statistical models were predicted between the IC50 of the colorectal cell line HCT to correlate the anticancer activity and the computed physicochemical and quantum chemical global reactivity descriptors. The predictive power of the models indicates a good correlation between the observed and the predicted activity.

Pimavanserin and Parkinson's Disease Psychosis: A Narrative Review.

Pimavanserin (PMV) is the first approved drug for treating hallucinations and delusions in Parkinson’s disease (PD) psychosis. Psychosis is one of the leading causes of nursing home placement in people with PD. Furthermore, hallucinations are a more frequent cause of institutionalization than motor disability or dementia related to PD. The management of PD psychosis involves antipsychotic medications. Most of the drugs in this class directly block dopamine D2 receptors, leading to significantly worsening motor symptoms in patients with PD. The most commonly used medications for managing PD psychosis are quetiapine, clozapine, and PMV. This literature review aims to study pimavanserin’s history, mechanism, clinical trials, and post-marketing experience. PMV is a potent 5-HT2A receptor antagonist/inverse agonist. Moreover, this drug can interact with 5-HT2C receptors. We calculated some physicochemical descriptors and pharmacokinetic properties of PMV. Eight clinical trials of PMV and PD psychosis are registered on ClinicalTrials.gov. Only four of them have complete results already published. Meta-analytic results showed that PMV efficacy is inferior to clozapine. However, PMV has a significantly lower number of side-effects for managing psychosis in PD. Medicare database assessment revealed 35% lower mortality with PMV compared to other atypical antipsychotics. Moreover, sensitive statistical analysis demonstrated that PMV is a protective factor for the risk of falls in individuals with PD.

Classification analysis of fatty acid synthase inhibitors using multialgorithms on topological descriptors and structural fingerprints.

Fatty acid synthase (FASN) is one of the enzymes required for fatty acid biosynthesis and is expressed as low or absent in most normal cells/tissues. However, this enzyme is upregulated in various cancer cells; hence, it can act as an important target to design and develop novel FASN inhibitors for cancer therapy. In the present investigation, a series of structurally diverse compounds that possessed FASN inhibitory activities were subjected to classification analysis using different algorithms such as support vector machine, decision tree, Naïve Bayes and random forest. The physicochemical descriptors and MACCS fingerprints were calculated using PaDEL software, and the WEKA software was utilized for the classification model building. The statistical parameters/confusion matrix calculated from the analysis revealed that the selected models have significant predictive performances. The results showed that the topological properties of the molecules are the main determinant for the activity classification. The key descriptors comprised of hydrogen bonding groups, especially acceptor (nHBAcc, minHBint9, minHBint5 and nwHBa), charge on the topological surface of the molecules (JGI10 & GGI2), ionization potential (GATS5i and GATS1i) and branching and distance between the groups (ETA_Eta_B_RC) are significantly contributed in the classification models. Further, the presence of heteroatoms (MACCSFP82, MACCSFP93 and MACCSFP131), especially nitrogen atom(s) and hydrogen bond acceptor groups (N-N group, NC(=O)N, N-C(=O)), actively contributed to the inhibitory activities. The results concluded that the topological polar properties concentrated in a specific region have significant FASN inhibitory activity. Hence, these results shall be used to develop novel molecules with increased FASN inhibitory activity.

Evaluating structure-based activity in a high-throughput assay for steroid biosynthesis

Data from a high-throughput human adrenocortical carcinoma assay (HT-H295R) for steroid hormone biosynthesis are available for > 2000 chemicals in single concentration and 654 chemicals in multi-concentration (mc). Previously, a metric describing the effect size of a chemical on the biosynthesis of 11 hormones was derived using mc data referred to as the maximum mean Mahalanobis distance (maxmMd). However, mc HT-H295R assay data remain unavailable for many chemicals. This work leverages existing HT-H295R assay data by constructing structure–activity relationships to make predictions for data-poor chemicals, including: (1) identification of individual structural descriptors, known as ToxPrint chemotypes, associated with increased odds of affecting estrogen or androgen synthesis; (2) a random forest (RF) classifier using physicochemical property descriptors to predict HT-H295R maxmMd binary (positive or negative) outcomes; and, (3) a local approach to predict maxmMd binary outcomes using nearest neighbors (NNs) based on two types of chemical fingerprints (chemotype or Morgan). Individual chemotypes demonstrated high specificity (85–98 %) for modulators of estrogen and androgen synthesis but with low sensitivity. The best RF model for maxmMd classification included 13 predicted physicochemical descriptors, yielding a balanced accuracy (BA) of 71 % with only modest improvement when hundreds of structural features were added. The best two NN models for binary maxmMd prediction demonstrated BAs of 85 and 81 % using chemotype and Morgan fingerprints, respectively. Using an external test set of 6302 chemicals (lacking HT-H295R data), 1241 were identified as putative estrogen and androgen modulators. Combined results across the three classification models (global RF model and two local NN models) predict that 1033 of the 6302 chemicals would be more likely to affect HT-H295R bioactivity. Together, these in silico approaches can efficiently prioritize thousands of untested chemicals for screening to further evaluate their effects on steroid biosynthesis.

Predicting mosquito repellents for clothing application from molecular fingerprint-based artificial neural network SAR models

ABSTRACT Spraying repellents on clothing limits toxicity and allergy problems that can occur when the repellents are directly applied to skin. This also allows the use of higher doses to ensure longer lasting effects. As the number of repellents available on the market is limited, it is necessary to propose new ones, especially by using in silico methods that reduce costs and time. In this context SAR models were built from a dataset of 2027 chemicals for which repellent activity on clothing was measured against Aedes aegypti. The interest of using either the ECFP or MACCS fingerprints as input neurons of a three-layer perceptron was evaluated. Transformation of MACCS bit strings into disjunctive tables led to interesting results. Models obtained with both types of fingerprints were compared to a model including physicochemical and topological descriptors.

QSTR Modeling to Find Relevant DFT Descriptors Related to the Toxicity of Carbamates.

Compounds containing carbamate moieties and their derivatives can generate serious public health threats and environmental problems due their high potential toxicity. In this study, a quantitative structure–toxicity relationship (QSTR) model has been developed by using one hundred seventy-eight carbamate derivatives whose toxicities in rats (oral administration) have been evaluated. The QSRT model was rigorously validated by using either tested or untested compounds falling within the applicability domain of the model. A structure-based evaluation by docking from a series of carbamates with acetylcholinesterase (AChE) was carried out. The toxicity of carbamates was predicted using physicochemical, structural, and quantum molecular descriptors employing a DFT approach. A statistical treatment was developed; the QSRT model showed a determination coefficient (R2) and a leave-one-out coefficient (Q2LOO) of 0.6584 and 0.6289, respectively.

Comparative analysis of an anthraquinone and chalcone derivatives-based virtual combinatorial library. A cheminformatics “proof-of-concept” study

A Laplacian scoring algorithm for gene selection and the Gini coefficient to identify the genes whose expression varied least across a large set of samples were the state-of-the-art methods used here. These methods have not been trialed for their feasibility in cheminformatics. This was a maiden attempt to investigate a complete comparative analysis of an anthraquinone and chalcone derivatives-based virtual combinatorial library. This computational “proof-of-concept” study illustrated the combinatorial approach used to explain how the structure of the selected natural products (NPs) undergoes molecular diversity analysis. A virtual combinatorial library (1.6 M) based on 20 anthraquinones and 24 chalcones was enumerated. The resulting compounds were optimized to the near drug-likeness properties, and the physicochemical descriptors were calculated for all datasets including FDA, Non-FDA, and NPs from ZINC 15. UMAP and PCA were applied to compare and represent the chemical space coverage of each dataset. Subsequently, the Laplacian score and Gini coefficient were applied to delineate feature selection and selectivity among properties, respectively. Finally, we demonstrated the diversity between the datasets by employing Murcko's and the central scaffolds systems, calculating three fingerprint descriptors and analyzing their diversity by PCA and SOM. The optimized enumeration resulted in 1,610,268 compounds with NP-Likeness, and synthetic feasibility mean scores close to FDA, Non-FDA, and NPs datasets. The overlap between the chemical space of the 1.6 M database was more prominent than with the NPs dataset. A Laplacian score prioritized NP-likeness and hydrogen bond acceptor properties (1.0 and 0.923), respectively, while the Gini coefficient showed that all properties have selective effects on datasets (0.81–0.93). Scaffold and fingerprint diversity indicated that the descending order for the tested datasets was FDA, Non-FDA, NPs and 1.6 M. Virtual combinatorial libraries based on NPs can be considered as a source of the combinatorial compound with NP-likeness properties. Furthermore, measuring molecular diversity is supposed to be performed by different methods to allow for comparison and better judgment.

3-[(1H-Benzo[d][1,2,3]triazol-1-yl)oxy]propyl 9-hydroxy-5a,5b,8,8,11a-pentamethyl-1-(prop-1-en-2-yl)icosahydro-3aH-cyclopenta[a]chrysene-3a-carboxylate

We herein report on the synthesis of a pentacyclic triterpene functionalized through derivation of betulinic acid with hydroxybenzotriazole. The compound was fully characterized by proton (1H-NMR), carbon-13 (13C-NMR), heteronuclear single quantum coherence (HSQC) and distortionless enhancement by polarization transfer (DEPT-135 and DEPT-90) nuclear magnetic resonance. Ultraviolet (UV), and Fourier-transform infrared (FTIR) spectroscopies as well as and high-resolution mass spectrometry (HRMS) were also adopted. Computational studies were conducted to foresee the interactions between compound 3 and phosphodiesterase 9, a relevant target in the field of neurodegenerative diseases. Additionally, preliminary calculation of physico-chemical descriptors was performed to evaluate the drug-likeness of compound 3.

Development of Novel Therapeutics for Schizophrenia Treatment Based on a Selective Positive Allosteric Modulation of α1-Containing GABAARs-In Silico Approach.

For the development of atypical antipsychotics, the selective positive allosteric modulation of the ionotropic GABAA receptor (GABAAR) has emerged as a promising approach. In the presented research, two unrelated methods were used for the development of QSAR models for selective positive allosteric modulation of 1-containing GABAARs with derivatives of imidazo [1,2-a]-pyridine. The development of conformation-independent QSAR models, based on descriptors derived from local molecular graph invariants and SMILES notation, was achieved with the Monte Carlo optimization method. From the vast pool of 0D, 1D, and 2D molecule descriptors, the GA-MLR method developed additional QSAR models. Various statistical methods were utilised for the determination of the developed models’ robustness, predictability, and overall quality, and according to the obtained results, all QSAR models are considered good. The molecular fragments that have a positive or negative impact on the studied activity were obtained from the studied molecules’ SMILES notations, and according to the obtained results, nine novel compounds were designed. The binding affinities to GABAAR of designed compounds were assessed with the application of molecular docking studies and the obtained results showed a high correlation with results obtained from QSAR modeling. To assess all designed molecules’ “drug-likeness”, their physicochemical descriptors were computed and utilised for the prediction of medicinal chemistry friendliness, pharmacokinetic properties, ADME parameters, and druglike nature.

Physicochemical analysis of Ugandan tea (Camellia sinensis) germplasm reveals potential commercial green and black tea varieties

ABSTRACT Tea (Camellia sinensis L.) is an important beverage consumed worldwide. In Uganda, it is the second-largest and highly prioritized export crop after coffee and provides the much-needed jobs to more than 800,000 people. Despite its importance in the country, the crop has received very limited research attention for its improvement and optimal utilization. This study was aimed at characterizing existing tea germplasm in Uganda to inform future breeding initiatives for market-preferred tea varieties. Fifty-eight advanced tea clones were randomly selected from the tea germplasm conserved at Uganda’s Tea Research Center in Rwebitaba and analyzed in the laboratory for eight physicochemical descriptors. Hierarchical cluster analysis performed on the 58 clones using Cluster-R package revealed four main clusters, indicating the existence of variation for physicochemical parameters within tested germplasm. This variation can be exploited to select among and/or improve the studied germplasm genetically for quality. High fermentation rate, which is one of the key indicators for good-quality black teas, revealed 15 fast-fermenter tea clones. Clone “UTR12/12” was identified as the most rapid fermenter, fully fermenting within 30 min, which implies a good clone for black tea. The clone also had the highest polyphenol content (26.7%), higher than the high-quality clone “UTR6/8” (control). Other tea clones whose polyphenol content was within the range of the control black tea clone were: “UTR144/10” (20.9%) and ‘UTR144/17, (20.66%). The identified promising black tea clones can be advanced to multi-location trials for further evaluation and selection for eventual variety release as commercial black-tea clones.

Insight into the Structure-Odor Relationship of Molecules: A Computational Study Based on Deep Learning.

Molecules with pleasant odors, unacceptable odors, and even serious toxicity are closely related to human social life. It is impractical to identify the odors of molecules in large quantities (particularly hazardous odors) using experimental methods. Computer-aided methods have currently attracted increasing attention for the prediction of molecular odors. Here, through models based on multilayer perceptron (MLP) and physicochemical descriptors (MLP-Des), MLP and molecular fingerprint, and convolutional neural network (CNN), we conduct the two-class prediction of odor/no odor, fruity/no odor, floral/no odor, and woody/no odor, and the multi-class prediction of fruity/flowery/woody/no odor on our newly refined molecular odor datasets. We show that three kinds of predictors can robustly predict molecular odors. The MLP-Des model not only exhibits the best prediction results (the AUC values are 0.99 and 0.86 for the two- and multi-classification models, respectively) but can also well reflect the characteristics of the structure–odor relationship of molecules. The CNN model takes 2D molecular images as input and can automatically extract the structural features related to molecular odors. The proposed models are of great help for the prediction of molecular odorants, understanding the underlying relationship between chemical structure and odor perception, and the discovery of new odorous and/or hazardous molecules.

QSAR, molecular docking and ADMET studies of quinoline, isoquinoline and quinazoline derivatives against Plasmodium falciparum malaria

With the aim of researching new antimalarial drugs, a series of quinoline, isoquinoline, and quinazoline derivatives were studied against the Plasmodium falciparum CQ-sensitive and MQ-resistant strain 3D7 protozoan parasite. DFT with B3LYP functional and 6-311G basis set was used to calculate quantum chemical descriptors for QSAR models. The molecular mechanics (MM2) method was used to calculate constitutional, physicochemical, and topological descriptors. By randomly dividing the dataset into training and test sets, we were able to construct reliable models using linear regression (MLR), nonlinear regression (MNLR), and artificial neural networks (ANN). The determination coefficient values indicate the predictive quality of the established models. The robustness and predictive power of the generated models were also confirmed via internal validation, external validation, the Y-randomization test, and the applicability domain. Furthermore, molecular docking studies were conducted to identify the key interactions between the studied molecules and the PfPMT receptor’s active site. The findings of this contribution study indicate that the antimalarial activity of these compounds against Plasmodium falciparum appears to be largely determined by four descriptors, i.e., total connectivity (Tcon), percentage of carbon (C (%)), density (D), and bond length between the two nitrogen atoms (Bond N–N). On the basis of the reliable QSAR model and molecular docking results, several new antimalarial compounds have been designed. The selection of drug candidates was performed according to drug-likeness and ADMET parameters.

Multifunctional Analysis of Chia Seed (Salvia hispanica L.) Bioactive Peptides Using Peptidomics and Molecular Dynamics Simulations Approaches

Chia seed peptides (CSP) can be a source of multifunctional biopeptides to treat non-communicable diseases. However, interactions and binding affinity involved in targeting specific receptors remains unexplored. In this study, molecular simulation techniques were used as virtual screening of CSP to determine drug-like candidates using a multi-target-directed ligand approach. CSP fraction with the best bioactivities in vitro was sequenced. Then, a prediction model was built using physicochemical descriptors (hydrophobicity, hydrophilicity, intestinal stability, antiangiogenic, antihypertensive, and anti-inflammatory) to calculate potential scores and rank possible biopeptides. Furthermore, molecular dynamics simulations (MDS) and ensemble molecular docking analysis were carried out using four human protein targets (ACE, angiotensin converting enzyme; VEGF, vascular endothelial growth factor; GLUC, glucocorticoid and MINC, mineralocorticoid receptors). Five known-sequence peptides (NNVFYPF, FNIVFPG, SRPWPIDY, QLQRWFR, GSRFDWTR) and five de novo peptides (DFKF, DLRF, FKAF, FRSF, QFRF) had the lowest energy score and higher affinity for ACE and VEGF. The therapeutic effects of these selected peptides can be related to the inhibition of the enzymes involved in angiogenesis and hypertension, due to formation of stable complexes with VEGF and ACE binding sites, respectively. The application of MDS is a good resource for identifying bioactive peptides for future experimental validation.

Urea derivatives carrying a thiophenylthiazole moiety: Design, synthesis, and evaluation of antitubercular and InhA inhibitory activities.

The recent emergence of drug-resistant strains of Mycobacterium tuberculosis (Mtb) has complicated and significantly slowed efforts to eradicate and/or reduce the worldwide incidence of life-threatening acute and chronic cases of tuberculosis. To overcome this setback, researchers have increased the intensity of their workto identify new small-molecule compounds that are expected to remain efficacious antimicrobials against Mtb. Here, we describe our effort to apply the principles of molecular hybridization to synthesize 16 compounds carrying thiophene and thiazole rings beside the core urea functionality (TTU1-TTU16). Following extensive structural characterization, the obtained compounds were initially evaluated for their antimycobacterial activity against Mtb H37Rv. Subsequently, three derivatives standing out with their anti-Mtb activity profiles and low cytotoxicity (TTU5, TTU6, and TTU12) were tested on isoniazid-resistant clinical isolates carrying katG and inhA mutations. Additionally, due to their pharmacophore similarities to the well-known InhA inhibitors, the molecules were screened for their enoyl acyl carrier protein reductase (InhA) inhibitory potentials. Molecular docking studies were performed to support the experimental enzyme inhibition data. Finally, drug-likeness of the selected compounds was established by theoretical calculations of physicochemical descriptors.

Advancing Rare-Earth Separation by Machine Learning.

Constituting the bulk of rare-earth elements, lanthanides need to be separated to fully realize their potential as critical materials in many important technologies. The discovery of new ligands for improving rare-earth separations by solvent extraction, the most practical rare-earth separation process, is still largely based on trial and error, a low-throughput and inefficient approach. A predictive model that allows high-throughput screening of ligands is needed to identify suitable ligands to achieve enhanced separation performance. Here, we show that deep neural networks, trained on the available experimental data, can be used to predict accurate distribution coefficients for solvent extraction of lanthanide ions, thereby opening the door to high-throughput screening of ligands for rare-earth separations. One innovative approach that we employed is a combined representation of ligands with both molecular physicochemical descriptors and atomic extended-connectivity fingerprints, which greatly boosts the accuracy of the trained model. More importantly, we synthesized four new ligands and found that the predicted distribution coefficients from our trained machine-learning model match well with the measured values. Therefore, our machine-learning approach paves the way for accelerating the discovery of new ligands for rare-earth separations.

QSAR modelling, molecular docking studies and ADMET predictions of polysubstituted pyridinylimidazoles as dual inhibitors of JNK3 and p38α MAPK

This research paper employs different types of QSAR modeling for the molecules that act as dual inhibitors of JNK3 and p38α MAPK. The SMILES notation and the local molecular graph invariants served as descriptors for QSAR model building, while the Monte Carlo optimization method was assigned the purpose of a model developer. GA-MLR was utilized to obtain a QSAR model from the pool of vast 2D molecule descriptors. To assess the developed models’ quality, robustness, and predictability, several statistical methods were employed, and these yielded favorable results. The molecular fragments which were responsible for the increase/decrease of the examined activity were identified and used for the computer-aided design of new compounds. Molecular docking studies were utilized for the final assessment of the designed inhibitors, highlighting an exceptional correlation with the QSAR modeling results. ADME parameters, pharmacokinetic properties, the drug-like nature, and medicinal chemistry friendliness were predicted by computing physicochemical descriptors to support drug discovery. Based on the obtained results, all the designed molecules possess high drug-likeness.

Carcinogenicity prediction using the index of ideality of correlation

ABSTRACT Carcinogenicity testing is necessary to protect human health and comply with regulations, but testing it with the traditionally used two-year rodent studies is time-consuming and expensive. In certain cases, such as for impurities, alternative methods may be convenient. Thus there is an urgent need for alternative approaches for reliable and robust assessments of carcinogenicity. The Monte Carlo technique with CORAL software is a tool to tackle this task for unknown compounds using available experimental data for a representative set of compounds. The models can be constructed with the simplified molecular input line entry system without additional physicochemical descriptors. We describe here a model based on a data set of 1167 substances. Matthew’s correlation coefficient values for calibration and validation sets are 0.747 and 0.577, respectively. Double bonds between carbon atoms and double bonds of oxygen atoms are the molecular features that indicate the carcinogenic potential of a compound.

QSAR, ADMET In Silico Pharmacokinetics, Molecular Docking and Molecular Dynamics Studies of Novel Bicyclo (Aryl Methyl) Benzamides as Potent GlyT1 Inhibitors for the Treatment of Schizophrenia.

Forty-four bicyclo ((aryl) methyl) benzamides, acting as glycine transporter type 1 (GlyT1) inhibitors, are developed using molecular modeling techniques. QSAR models generated by multiple linear and non-linear regressions affirm that the biological inhibitory activity against the schizophrenia disease is strongly and significantly correlated with physicochemical, geometrical and topological descriptors, in particular: Hydrogen bond donor, polarizability, surface tension, stretch and torsion energies and topological diameter. According to in silico ADMET properties, the most active ligands (L6, L9, L30, L31 and L37) are the molecules having the highest probability of penetrating the central nervous system (CNS), but the molecule 32 has the highest probability of being absorbed by the gastrointestinal tract. Molecular docking results indicate that Tyr124, Phe43, Phe325, Asp46, Phe319 and Val120 amino acids are the active sites of the dopamine transporter (DAT) membrane protein, in which the most active ligands can inhibit the glycine transporter type 1 (GlyT1). The results of molecular dynamics (MD) simulation revealed that all five inhibitors remained stable in the active sites of the DAT protein during 100 ns, demonstrating their promising role as candidate drugs for the treatment of schizophrenia.

Bioaccumulation and bioamplification of pharmaceuticals and endocrine disruptors in aquatic insects

Environmental fate of emerging contaminants such as pharmaceuticals and endocrine disrupting compounds at the aquatic terrestrial boundary are largely unexplored. Aquatic insects connect aquatic and terrestrial food webs as their life cycle includes aquatic and terrestrial life stages, thus they represent an important inter-habitat linkage not only for energy and nutrient flow, but also for contaminant transfer to terrestrial environments. We measured the concentrations of pharmaceuticals and endocrine disrupting compounds in the larval and adult tissues (last larval stages and teneral adults) of five Odonata species sampled in a wastewater-impacted river, in order to examine their bioaccumulation and bioamplification at different taxonomic levels. Twenty different compounds were bioaccumulated in insect tissues, with majority having higher concentrations (up to 90% higher) in aquatic larvae compared to terrestrial adults (reaching 88 ng/g for 1H-benzotriazole). However, increased concentration in adults was observed for seven compounds in at least one suborder (41% of the accumulated), confirming contaminants bioamplification across the metamorphosis. Both, bioaccumulation and bioamplification differed at various taxa levels; the order (Odonata), suborder (Anisoptera and Zygoptera) and species level. Highest variability was observed between Anisoptera and Zygoptera, due to the underlying differences in their ecology. Generally, Zygoptera had higher concentrations of contaminants in both larvae and adults. Additionally, we aimed at predicting effects of contaminant properties on bioaccumulation and bioamplification patterns using the commonly used physicochemical and pharmacokinetic descriptors on both order and suborder levels, however, neither of the two processes could be consistently predicted with simple linear models. Our study highlights the importance of taxonomy in studies aiming at advancing the understanding of contaminant exchange between aquatic and terrestrial food webs, as higher taxonomic categories include ecologically diverse groups, whose contribution to “the dark side of subsidies” could substantially differ.