Cheminformatics approach to identify andrographolide derivatives as dual inhibitors of methyltransferases (nsp14 and nsp16) of SARS-CoV-2.

The Covid-19 pandemic outbreak has accelerated tremendous efforts to discover a therapeutic strategy that targets severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) to control viral infection. Various viral proteins have been identified as potential drug targets, however, to date, no specific therapeutic cure is available against the SARS-CoV-2. To address this issue, the present work reports a systematic cheminformatic approach to identify the potent andrographolide derivatives that can target methyltransferases of SARS-CoV-2, i.e. nsp14 and nsp16 which are crucial for the replication of the virus and host immune evasion. A consensus of cheminformatics methodologies including virtual screening, molecular docking, ADMET profiling, molecular dynamics simulations, free-energy landscape analysis, molecular mechanics generalized born surface area (MM-GBSA), and density functional theory (DFT) was utilized. Our study reveals two new andrographolide derivatives (PubChem CID: 2734589 and 138968421) as natural bioactive molecules that can form stable complexes with both proteins via hydrophobic interactions, hydrogen bonds and electrostatic interactions. The toxicity analysis predicts class four toxicity for both compounds with LD50 value in the range of 500-700 mg/kg. MD simulation reveals the stable formation of the complex for both the compounds and their average trajectory values were found to be lower than the control inhibitor and protein alone. MMGBSA analysis corroborates the MD simulation result and showed the lowest energy for the compounds 2734589 and 138968421. The DFT and MEP analysis also predicts the better reactivity and stability of both the hit compounds. Overall, both andrographolide derivatives exhibit good potential as potent inhibitors for both nsp14 and nsp16 proteins, however, in-vitro and in vivo assessment would be required to prove their efficacy and safety in clinical settings. Moreover, the drug discovery strategy aiming at the dual target approach might serve as a useful model for inventing novel drug molecules for various other diseases.

Enhancing the Small-Scale Screenable Biological Space beyond Known Chemogenomics Libraries with Gray Chemical Matter─Compounds with Novel Mechanisms from High-Throughput Screening Profiles.

Phenotypic assays have become an established approach to drug discovery. Greater disease relevance is often achieved through cellular models with increased complexity and more detailed readouts, such as gene expression or advanced imaging. However, the intricate nature and cost of these assays impose limitations on their screening capacity, often restricting screens to well-characterized small compound sets such as chemogenomics libraries. Here, we outline a cheminformatics approach to identify a small set of compounds with likely novel mechanisms of action (MoAs), expanding the MoA search space for throughput limited phenotypic assays. Our approach is based on mining existing large-scale, phenotypic high-throughput screening (HTS) data. It enables the identification of chemotypes that exhibit selectivity across multiple cell-based assays, which are characterized by persistent and broad structure activity relationships (SAR). We validate the effectiveness of our approach in broad cellular profiling assays (Cell Painting, DRUG-seq, and Promotor Signature Profiling) and chemical proteomics experiments. These experiments revealed that the compounds behave similarly to known chemogenetic libraries, but with a notable bias toward novel protein targets. To foster collaboration and advance research in this area, we have curated a public set of such compounds based on the PubChem BioAssay dataset and made it available for use by the scientific community.

RDCanon: A Python Package for Canonicalizing the Order of Tokens in SMARTS Queries.

SMARTS is a widely used language in cheminformatics for defining substructural queries for database lookups, reaction templates for chemical transformations, and other applications. As an extension to SMILES, many SMARTS patterns can represent the same query. Despite this, no canonicalization algorithm invariant of the line notation sequence or atomic numbering is publicly available. Here, we introduce RDCanon, an open-source Python package that can be used to standardize SMARTS queries. RDCanon is designed to ensure that the sequence of atomic queries remains consistent for all graphs representing the same substructure query and to ensure a canonical sequence of primitives within each individual atom query; furthermore, the algorithm can be applied to canonicalize the order of reactants, agents, and products and their atom map numbers in reaction SMARTS templates. As part of its canonicalization algorithm, RDCanon provides a mechanism in which the canonicalized SMARTS is optimized for speed against specific molecular databases. Several case studies are provided to showcase improved efficiency in substructure matching and retrosynthetic analysis.

PaCh (Packed Chemicals): Computationally Effective Binary Format for Chemical Structure Encoding.

In this work, we propose a versatile molecule and reaction encoding binary data format that aims to bridge the gap between the advantages of SMILES, like local stereo- and implicit hydrogen encoding, and block-structured MDL MOL with a 2D layout and explicit bond encoding, while addressing their respective limitations. Our new format introduces a balance between size efficiency, processing speed, and comprehensive representation, making it well-suited for various applications in cheminformatics, including deep learning, data storage, and searching. By offering an explicit approach to store atom connectivity (including implicit hydrogens), electronic state, stereochemistry, and other crucial molecular attributes, our proposal seeks to enhance data storage efficiency and promote interoperability among different software tools.

Flexible scaffold-based cheminformatics approach for polypharmacological drug design.

Effective treatments for complex central nervous system (CNS) disorders require drugs with polypharmacology and multifunctionality, yet designing such drugs remains a challenge. Here, we present a flexible scaffold-based cheminformatics approach (FSCA) for the rational design of polypharmacological drugs. FSCA involves fitting a flexible scaffold to different receptors using different binding poses, as exemplified by IHCH-7179, which adopted a "bending-down" binding pose at 5-HT2AR to act as an antagonist and a "stretching-up" binding pose at 5-HT1AR to function as an agonist. IHCH-7179 demonstrated promising results in alleviating cognitive deficits and psychoactive symptoms in mice by blocking 5-HT2AR for psychoactive symptoms and activating 5-HT1AR to alleviate cognitive deficits. By analyzing aminergic receptor structures, we identified two featured motifs, the "agonist filter" and "conformation shaper," which determine ligand binding pose and predict activity at aminergic receptors. With these motifs, FSCA can be applied to the design of polypharmacological ligands at other receptors.

Pan-Cancer Prediction of Cell-Line Drug Sensitivity Using Network-Based Methods.

The development of reliable predictive models for individual cancer cell lines to identify an optimal cancer drug is a crucial step to accelerate personalized medicine, but vast differences in cancer cell lines and drug characteristics make it quite challenging to develop predictive models that result in high predictive power and explain the similarity of cell lines or drugs. Our study proposes a novel network-based methodology that breaks the problem into smaller, more interpretable problems to improve the predictive power of anti-cancer drug responses in cell lines. For the drug-sensitivity study, we used the GDSC database for 915 cell lines and 200 drugs. The theory of optimal mass transport was first used to separately cluster cell lines and drugs, using gene-expression profiles and extensive cheminformatic drug features, represented in a form of data networks. To predict cell-line specific drug responses, random forest regression modeling was separately performed for each cell-line drug cluster pair. Post-modeling biological analysis was further performed to identify potential biological correlates associated with drug responses. The network-based clustering method resulted in 30 distinct cell-line drug cluster pairs. Predictive modeling on each cell-line-drug cluster outperformed alternative computational methods in predicting drug responses. We found that among the four drugs top-ranked with respect to prediction performance, three targeted the PI3K/mTOR signaling pathway. Predictive modeling on clustered subsets of cell lines and drugs improved the prediction accuracy of cell-line specific drug responses. Post-modeling analysis identified plausible biological processes associated with drug responses.

Discrimination of Adulterated Ginkgo Biloba Products Based on 2T2D Correlation Spectroscopy in UV-Vis Range.

Ginkgo biloba is a popular medicinal plant widely used in numerous herbal products, including food supplements. Due to its popularity and growing economic value, G. biloba leaf extract has become the target of economically motivated adulterations. There are many reports about the poor quality of ginkgo products and their adulteration, mainly by adding flavonols, flavonol glycosides, or extracts from other plants. In this work, we developed an approach using two-trace two-dimensional correlation spectroscopy (2T2D COS) in UV-Vis range combined with multilinear principal component analysis (MPCA) to detect potential adulteration of twenty G. biloba food supplements. UV-Vis spectral data are obtained for 80% methanol and aqueous extracts in the range of 245-410 nm. Three series of two-dimensional correlation spectra were interpreted by visual inspection and using MPCA. The proposed relatively quick and straightforward approach successfully differentiated supplements adulterated with rutin or those lacking ginkgo leaf extract. Supporting information about adulteration was obtained from the difference between the DPPH radical scavenging capacity of both extracts and from chromatographic (HPLC-DAD) fingerprints of methanolic samples.

Computational Screening for the Anticancer Potential of Seed-Derived Antioxidant Peptides: A Cheminformatic Approach.

Some seed-derived antioxidant peptides are known to regulate cellular modulators of ROS production, including those proposed to be promising targets of anticancer therapy. Nevertheless, research in this direction is relatively slow owing to the inevitable time-consuming nature of wet-lab experimentations. To help expedite such explorations, we performed structure-based virtual screening on seed-derived antioxidant peptides in the literature for anticancer potential. The ability of the peptides to interact with myeloperoxidase, xanthine oxidase, Keap1, and p47phox was examined. We generated a virtual library of 677 peptides based on a database and literature search. Screening for anticancer potential, non-toxicity, non-allergenicity, non-hemolyticity narrowed down the collection to five candidates. Molecular docking found LYSPH as the most promising in targeting myeloperoxidase, xanthine oxidase, and Keap1, whereas PSYLNTPLL was the best candidate to bind stably to key residues in p47phox. Stability of the four peptide-target complexes was supported by molecular dynamics simulation. LYSPH and PSYLNTPLL were predicted to have cell- and blood-brain barrier penetrating potential, although intolerant to gastrointestinal digestion. Computational alanine scanning found tyrosine residues in both peptides as crucial to stable binding to the targets. Overall, LYSPH and PSYLNTPLL are two potential anticancer peptides that deserve deeper exploration in future.

Unsupervised Representation Learning for Proteochemometric Modeling.

In silico protein-ligand binding prediction is an ongoing area of research in computational chemistry and machine learning based drug discovery, as an accurate predictive model could greatly reduce the time and resources necessary for the detection and prioritization of possible drug candidates. Proteochemometric modeling (PCM) attempts to create an accurate model of the protein-ligand interaction space by combining explicit protein and ligand descriptors. This requires the creation of information-rich, uniform and computer interpretable representations of proteins and ligands. Previous studies in PCM modeling rely on pre-defined, handcrafted feature extraction methods, and many methods use protein descriptors that require alignment or are otherwise specific to a particular group of related proteins. However, recent advances in representation learning have shown that unsupervised machine learning can be used to generate embeddings that outperform complex, human-engineered representations. Several different embedding methods for proteins and molecules have been developed based on various language-modeling methods. Here, we demonstrate the utility of these unsupervised representations and compare three protein embeddings and two compound embeddings in a fair manner. We evaluate performance on various splits of a benchmark dataset, as well as on an internal dataset of protein-ligand binding activities and find that unsupervised-learned representations significantly outperform handcrafted representations.

Systems Biology and Chemoinformatics-Based Strategies to Explore the Biological Mechanism of Fugui Wenyang Decoction in Treating Vascular Dementia Rats.

OBJECTIVE: To explore the biological mechanism of Fugui Wenyang Decoction (FGWYD) in treating vascular dementia (VD) rats based on systems pharmacology, proteomics, and a multidirectional pharmacology integration strategy. METHODS: Chemoinformatics was utilized to construct and analyze the FGWYD-VD protein-protein interaction (PPI) network. Then, the total protein in the brain tissue of the infarcted side of the rat was extracted for protein identification, pattern identification, and protein quantitative analysis. The differentially expressed proteins are analyzed by bioinformatics. Finally, the important proteins in the oxidative stress-related biological process proteins and indicators were detected through experimental pharmacology to verify the findings of systems biology and chemoinformatics. RESULTS: There were a total of 73 FGWYD components with 245 FGWYD and 145 VD genes. The results of GO enrichment analysis and pathway enrichment analysis showed that MBHD may regulate the inflammation module, oxidative stress, the synaptic plasticity regulation module, and the neuronal apoptosis section module. Compared with the sham operation group, there were 23 upregulated proteins and 17 downregulated proteins in the model group (P < 0.05). Compared with the model group, there were 16 upregulated proteins and 10 downregulated proteins in the FGWYD group (P < 0.05). Bioinformatics analysis shows that those proteins were closely related to processes such as inflammation, oxidative stress, neuronal apoptosis, neuronal growth and differentiation, signaling pathways, and transcriptional regulation. Multidirectional pharmacology further verified the neuroprotective mechanism of the Nrf2/HO-1 pathway in FGWYD treatment of VD. CONCLUSION: The mechanism of FGWYD in the treatment of VD may be related to inflammation, oxidative stress, angiogenesis, and neuronal apoptosis.

An integrated molecular modeling protocol for drug screening based on conceptual density functional theory and chemoinformatics for the study of marine cyclopeptides.

An integrated molecular modeling protocol resulting from the combination of conceptual density functional theory (CDFT) chemical reactivity descriptors with several chemoinformatics tools has been used for the study of the chemical reactivity and bioactivity properties of a group of marine cyclic peptides. CP-CDFT is a branch of computational chemistry and molecular modeling dedicated to the study of peptides. The protocol allowed the estimation of the CDFT-based reactivity indices together with the associated physicochemical parameters that can help to identify the ability of the studied peptides to behave as potential useful drugs. This was complemented with an analysis of the bioactivity and pharmacokinetics parameters related to the ADMET (absorption, distribution, metabolism, excretion, and toxicity) features. Some examples related to the ability of the CDFT-based chemical reactivity descriptors for the prediction of the pKas of the peptides as well as their potential as AGE inhibitors are also presented.

An Integrated Mutually Oriented "Chemical Profiling-Pharmaceutical Effect" Strategy for Screening Discriminating Markers of Underlying Hepatoprotective Effects to Distinguish Garden-Cultivated from Mountain-Cultivated Ginseng.

Garden-cultivated Ginseng (GG) and mountain-cultivated Ginseng (MG) both belong to Panax Ginseng C. A. Meyer. However, the effective substances which can be used to distinguish GG from MG remain obscure. Therefore, the purpose of this study was to screen for discriminating markers that can assist in the correct identification of GG and MG. HPLC Q-TOF/MS and various chemometrics methods were used to analyze the chemical profiles of 13 batches of Ginseng and to explore the characteristic constituents of both GG and MG. The hepatocyte-protecting effects of GG and MG were investigated through a paclitaxel-induced liver injury model. Through a combination of correlation analysis and bioinformatic techniques, markers for differentiation between GG and MG were ascertained. A total of 40 and 41 compounds were identified in GG and MG, respectively, and 15 characteristic ingredients contributed significantly to the discrimination of GG from MG. Correlation analysis and network pharmacology were applied and ginsenosides Rg1, Re, Rb1, Rc, Rb2, and Rg3 were found to be discriminating markers of GG and MG. Six markers for the identification of GG and MG were screened out by a step-wise mutually oriented "chemical profiling-pharmaceutical effect" correlation strategy, which is of great significance for future quality assessment of Ginseng products.

Network Pharmacology Analysis of the Identification of Phytochemicals and Therapeutic Mechanisms of Paeoniae Radix Alba for the Treatment of Asthma.

BACKGROUND: Paeoniae Radix Alba (PRA), the root of the plant Paeonia lactiflora Pall., has been suggested to play an important role for the treatment of asthma. A biochemical understanding of the clinical effects of Paeoniae Radix Alba is needed. Here, we explore the phytochemicals and therapeutic mechanisms via a systematic and comprehensive network pharmacology analysis. METHODS: Through TCMSP, PubChem, GeneCards database, and SwissTargetPrediction online tools, potential targets of active ingredients from PRA for the treatment of asthma were obtained. Cytoscape 3.7.2 was used to determine the target of active ingredients of PRA. Target protein interaction (PPI) network was constructed through the STRING database. The Gene Ontology (GO) biological process and Kyoto Encyclopedia of Genes and Genes (KEGG) pathway enrichment analysis were analyzed through the biological information annotation database (DAVID). RESULTS: Our results indicate that PRA contains 21 candidate active ingredients with the potential to treat asthma. The enrichment analysis of GO and KEGG pathways found that the treatment of asthma by PRA may be related to the process of TNF (tumor necrosis factor) release, which can regulate and inhibit multiple signaling pathways such as ceramide signaling. CONCLUSIONS: Our work provides a phytochemical basis and therapeutic mechanisms of PRA for the treatment of asthma, which provides new insights on further research on PRA.

Mol2Context-vec: learning molecular representation from context awareness for drug discovery.

With the rapid development of proteomics and the rapid increase of target molecules for drug action, computer-aided drug design (CADD) has become a basic task in drug discovery. One of the key challenges in CADD is molecular representation. High-quality molecular expression with chemical intuition helps to promote many boundary problems of drug discovery. At present, molecular representation still faces several urgent problems, such as the polysemy of substructures and unsmooth information flow between atomic groups. In this research, we propose a deep contextualized Bi-LSTM architecture, Mol2Context-vec, which can integrate different levels of internal states to bring dynamic representations of molecular substructures. And the obtained molecular context representation can capture the interactions between any atomic groups, especially a pair of atomic groups that are topologically distant. Experiments show that Mol2Context-vec achieves state-of-the-art performance on multiple benchmark datasets. In addition, the visual interpretation of Mol2Context-vec is very close to the structural properties of chemical molecules as understood by humans. These advantages indicate that Mol2Context-vec can be used as a reliable and effective tool for molecular expression. Availability: The source code is available for download in https://github.com/lol88/Mol2Context-vec.

Synthesis of Indofulvin Pseudo-Natural Products Yields a New Autophagy Inhibitor Chemotype.

Chemical and biological limitations in bioactive compound design based on natural product (NP) structure can be overcome by the combination of NP-derived fragments in unprecedented arrangements to afford "pseudo-natural products" (pseudo-NPs). A new pseudo-NP design principle is described, i.e., the combination of NP-fragments by transformations that are not part of current biosynthesis pathways. A collection of indofulvin pseudo-NPs is obtained from 2-hydroxyethyl-indoles and ketones derived from the fragment-sized NP griseofulvin by means of an iso-oxa-Pictet-Spengler reaction. Cheminformatic analysis indicates that the indofulvins reside in an area of chemical space sparsely covered by NPs, drugs, and drug-like compounds and they may combine favorable properties of these compound classes. Biological evaluation of the compound collection in different cell-based assays and the unbiased high content cell painting assay reveal that the indofulvins define a new autophagy inhibitor chemotype that targets mitochondrial respiration.

UHPLC-MS-based metabolomics and chemoinformatics study reveals the neuroprotective effect and chemical characteristic in Parkinson's disease mice after oral administration of Wen-Shen-Yang-Gan decoction.

Parkinson's disease (PD), the typical neurodegenerative disease, is characterized by the progressive loss of dopaminergic neurons in the substantia nigra (SN). However, no therapeutic agent used currently could slow down neuronal cell loss so as to decelerate or halt the progression of PD. Traditional Chinese medicine (TCM) has been utilized to treat the dysfunction of the autonomic nervous system. Wen-Shen-Yang-Gan decoction (WSYGD) has a good effect on the clinical treatment of PD with constipation. However, it is not clear which ingredients and what mechanism are responsible for the therapeutic effect. In this study, the pharmacodynamic study of WSYGD in PD mice was applied. Concurrently, a novel method for the identification of metabolic profiles of WSYGD has been developed. Finally, we found that WSYGD could protect the PD mice induced by rotenone. The underlying mechanism of the protective effect may be related to the reduction of the DA neurons apoptosis via reducing inflammatory reaction. By virtue of UPLC-MS and chemoinformatics method, 35 prototype compounds and 27 metabolites were filtered out and tentatively characterized. In conclusion, this study provides an insight into the metabolism of WSYGD in vivo to enable understanding of the metabolic process and therapeutic mechanism of PD.

Chemoinformatics Analyses of Tau Ligands Reveal Key Molecular Requirements for the Identification of Potential Drug Candidates against Tauopathies.

Tau is a highly soluble protein mainly localized at a cytoplasmic level in the neuronal cells, which plays a crucial role in the regulation of microtubule dynamic stability. Recent studies have demonstrated that several factors, such as hyperphosphorylation or alterations of Tau metabolism, may contribute to the pathological accumulation of protein aggregates, which can result in neuronal death and the onset of a number of neurological disorders called Tauopathies. At present, there are no available therapeutic remedies able to reduce Tau aggregation, nor are there any structural clues or guidelines for the rational identification of compounds preventing the accumulation of protein aggregates. To help identify the structural properties required for anti-Tau aggregation activity, we performed extensive chemoinformatics analyses on a dataset of Tau ligands reported in ChEMBL. The performed analyses allowed us to identify a set of molecular properties that are in common between known active ligands. Moreover, extensive analyses of the fragment composition of reported ligands led to the identification of chemical moieties and fragment combinations prevalent in the more active compounds. Interestingly, many of these fragments were arranged in recurring frameworks, some of which were clearly present in compounds currently under clinical investigation. This work represents the first in-depth chemoinformatics study of the molecular properties, constituting fragments and similarity profiles, of known Tau aggregation inhibitors. The datasets of compounds employed for the analyses, the identified molecular fragments and their combinations are made publicly available as supplementary material.

Learning to SMILES: BAN-based strategies to improve latent representation learning from molecules.

Computational methods have become indispensable tools to accelerate the drug discovery process and alleviate the excessive dependence on time-consuming and labor-intensive experiments. Traditional feature-engineering approaches heavily rely on expert knowledge to devise useful features, which could be costly and sometimes biased. The emerging deep learning (DL) methods deliver a data-driven method to automatically learn expressive representations from complex raw data. Inspired by this, researchers have attempted to apply various deep neural network models to simplified molecular input line entry specification (SMILES) strings, which contain all the composition and structure information of molecules. However, current models usually suffer from the scarcity of labeled data. This results in a low generalization ability of SMILES-based DL models, which prevents them from competing with the state-of-the-art computational methods. In this study, we utilized the BiLSTM (bidirectional long short term merory) attention network (BAN) in which we employed a novel multi-step attention mechanism to facilitate the extracting of key features from the SMILES strings. Meanwhile, SMILES enumeration was utilized as a data augmentation method in the training phase to substantially increase the number of labeled data and enlarge the probability of mining more patterns from complex SMILES. We again took advantage of SMILES enumeration in the prediction phase to rectify model prediction bias and provide a more accurate prediction. Combined with the BAN model, our strategies can greatly improve the performance of latent features learned from SMILES strings. In 11 canonical absorption, distribution, metabolism, excretion and toxicity-related tasks, our method outperformed the state-of-the-art approaches.

Cheminformatic Profiling and Hit Prioritization of Natural Products with Activities against Methicillin-Resistant Staphylococcus aureus (MRSA).

Several natural products (NPs) have displayed varying in vitro activities against methicillin-resistant Staphylococcus aureus (MRSA). However, few of these compounds have not been developed into potential antimicrobial drug candidates. This may be due to the high cost and tedious and time-consuming process of conducting the necessary preclinical tests on these compounds. In this study, cheminformatic profiling was performed on 111 anti-MRSA NPs (AMNPs), using a few orally administered conventional drugs for MRSA (CDs) as reference, to identify compounds with prospects to become drug candidates. This was followed by prioritizing these hits and identifying the liabilities among the AMNPs for possible optimization. Cheminformatic profiling revealed that most of the AMNPs were within the required drug-like region of the investigated properties. For example, more than 76% of the AMNPs showed compliance with the Lipinski, Veber, and Egan predictive rules for oral absorption and permeability. About 34% of the AMNPs showed the prospect to penetrate the blood-brain barrier (BBB), an advantage over the CDs, which are generally non-permeant of BBB. The analysis of toxicity revealed that 59% of the AMNPs might have negligible or no toxicity risks. Structure-activity relationship (SAR) analysis revealed chemical groups that may be determinants of the reported bioactivity of the compounds. A hit prioritization strategy using a novel "desirability scoring function" was able to identify AMNPs with the desired drug-likeness. Hit optimization strategies implemented on AMNPs with poor desirability scores led to the design of two compounds with improved desirability scores.

In silico prediction of chemical-induced hematotoxicity with machine learning and deep learning methods.

Chemical-induced hematotoxicity is an important concern in the drug discovery, since it can often be fatal when it happens. It is quite useful for us to give special attention to chemicals which can cause hematotoxicity. In the present study, we focused on in silico prediction of chemical-induced hematotoxicity with machine learning (ML) and deep learning (DL) methods. We collected a large data set contained 632 hematotoxic chemicals and 1525 approved drugs without hematotoxicity. Computational models were built using several different machine learning and deep learning algorithms integrated on the Online Chemical Modeling Environment (OCHEM). Based on the three best individual models, a consensus model was developed. It yielded the prediction accuracy of 0.83 and balanced accuracy of 0.77 on external validation. The consensus model and the best individual model developed with random forest regression and classification algorithm (RFR) and QNPR descriptors were made available at https://ochem.eu/article/135149 , respectively. The relevance of 8 commonly used molecular properties and chemical-induced hematotoxicity was also investigated. Several molecular properties have an obvious differentiating effect on chemical-induced hematotoxicity. Besides, 12 structural alerts responsible for chemical hematotoxicity were identified using frequency analysis of substructures from Klekota-Roth fingerprint. These results should provide meaningful knowledge and useful tools for hematotoxicity evaluation in drug discovery and environmental risk assessment.