Predicting Cell Association of Surface-Modified Nanoparticles Using Protein Corona Structure – Activity Relationships (PCSAR)

2015-05-20 09.25.22 amNanoparticles are likely to interact in real-case application scenarios with mixtures of proteins and biomolecules that will absorb onto their surface forming the so-called protein corona. Information related to the composition of the protein corona and net cell association was collected from literature for a library of surface-modified gold and silver nanoparticles. For each protein in the corona, sequence information was extracted and used to calculate physicochemical properties and statistical descriptors. Data cleaning and preprocessing techniques including statistical analysis and feature selection methods were applied to remove highly correlated, redundant and non-significant features. A weighting technique was applied to construct specific signatures that represent the corona composition for each nanoparticle. Using this basic set of protein descriptors, a new Protein Corona Structure-Activity Relationship (PCSAR) that relates net cell association with the physicochemical de- scriptors of the proteins that form the corona was developed and validated. The features that resulted from the feature selec- tion were in line with already published literature, and the computational model constructed on these features had a good ac- curacy (R2LOO=0.76 and R2LMO(25%)=0.72) and stability, with the advantage that the fingerprints based on physicochemical de- scriptors were independent of the specific proteins that form the corona.

Kamath P, Fernández A, Giralt F, Rallo R. Predicting Cell Association of Surface-Modified Nanoparticles Using Protein Corona Structure – Activity Relationships (PCSAR). Current Topics in Medicinal Chemistry, 2015, (15)18


Use of Quasi-SMILES and Monte Carlo Optimization to Develop Quantitative Feature Property/Activity Relationships (QFPR/QFAR) for Nanomaterials

2015-05-20 09.20.39 amThe CORAL software ( has been used to develop quantitative feature–property/activity relationships (QFPRs/QFARs) for the prediction of endpoints related to different categories of nanomaterials. In contrast to previous models built up by using CORAL from a representation of the molecular structure by using simplified molecular input-line entry system (SMILES), the current QFPR/QFARs are based on an integrated representation of acting conditions (i.e., a combination of physicochemical and/or biochemical factors) of nanomaterials via the so-called quasi-SMILES notation. In contrast to traditional quantita- tive structure – property / activity relationships (QSPRs/QSARs), the new models are able to provide new insight on the conditions of acting of substances (e.g., chemicals and nanomaterials) independently of their molecular structure. The de- velopment and validation of the QFPR/QFAR models was carried out following the OECD principles. The statistical qual- ity of models developed from quasi-SMILES is acceptable, with values for the determination coefficient in the range of 0.70 to 0.85 for various endpoints of environmental and human health relevance. Perspectives of the QFPR/QFAR and their interaction and overlapping with traditional QSPR/QSAR are also discussed.

Toropov AA, Rallo R, Toropova AP. Use of Quasi-SMILES and Monte Carlo Optimization to Develop Quantitative Feature Property/Activity Relationships (QFPR/QFAR) for Nanomaterials. Current Topics in Medicinal Chemistry, 2015, (15)18

  • DOI: 10.2174/1568026615666150506152000
  • PUBMED: 25961527

Quantitative Structure-Activity Relationships for Cellular Uptake of Surface-Modified Nanoparticles

Workflow for QSAR development. Molecular descriptors were calculated via MOE using the SMILES representation of the surface-modifying organic molecules, followed by data pre-processing to prune descriptors having same values for >95% of the NPs. Both linear and nonlinear (ε-SVR) regressions were used for QSAR development along with SFS and SFFS for descriptor selection and bootstrapping based validation for prediction accuracy estimation. Once the desired QSARs were identified, robustness validation was carried out followed by applicability domain (AD) analysis.

Quantitative structure-activity relationships (QSARs) were developed, for cellular uptake of nanoparticles (NPs) of the same iron oxide core but with different surface-modifying organic molecules, based on linear and non-linear (epsilon support vector regression (ε-SVR)). A linear QSAR provided high prediction accuracy of R2=0.751 (coefficient of determination) using 11 descriptors selected from an initial pool of 184 descriptors calculated for the NP surface-modifying molecules, while a ε-SVR based QSAR with only 6 descriptors improved prediction accuracy to R2 = 0.806. The linear and ε-SVR based QSARs both demonstrated good robustness and well spanned applicability domains. It is suggested that the approach of evaluating pertinent descriptors and their significance, via QSAR analysis, to cellular NP uptake could support planning and interpretation of toxicity studies as well as provide guidance for the tailor-design NPs with respect to targeted cellular uptake for various applications.

Liu R, Rallo R, Bilal M, Cohen Y. Quantitative Structure-Activity Relationships for Cellular Uptake of Surface-Modified Nanoparticles. Combinatorial Chemistry and High Throughput Screening, 2015, (18)3

Optimal descriptor as a translator of eclectic data into prediction of cytotoxicity of metal oxide nanoparticles under different conditions

S01476513The Monte Carlo technique has been used to build up quantitative structure–activity relationships (QSARs) for prediction of dark cytotoxicity and photo-induced cytotoxicity of metal oxide nanoparticles to bacteriaEscherichia coli (minus logarithm of lethal concentration for 50% bacteria pLC50, LC50 in mol/L). The representation of nanoparticles include (i) in the case of the dark cytotoxicity a simplified molecular input-line entry system (SMILES), and (ii) in the case of photo-induced cytotoxicity a SMILES plus symbol ‘ ^ ’. The predictability of the approach is checked up with six random distributions of available data into the visible training and calibration sets, and invisible validation set. The statistical characteristics of these models are correlation coefficient 0.90–0.94 (training set) and 0.73–0.98 (validation set).

Toropova AP, Toropov AA, Rallo R, Leszczynska D, Leszczynski J. Optimal descriptor as a translator of eclectic data into prediction of cytotoxicity for metal oxide nanoparticles under different conditions. Ecotoxicology and Environmental Safety., 2015, 112, 39-45

  • DOI: 10.1016/j.ecoenv.2014.10.003

Fault Detection and Isolation in a Spiral-Wound Reverse Osmosis (RO) Desalination Plant

ie-2013-03603x_0008Sensor fault detection and isolation (SFDI) approaches, based on support vector regression (SVR) plant sensor models and self-organizing-map (SOM) analysis, were investigated for application to reverse osmosis (RO) desalination plant operation. SFDI-SVR and SFDI-SOM were assessed using operational data from a small spiral-wound RO pilot plant and synthetic faulty data generated as perturbations relative to normal plant operational data. SFDI-SVR was achieved without false negative (FN) detections for sensor deviations of |10%| and FN detections of, at the most, |5%|, and for sensor deviations of |4%| at sensor fault detection (FD) thresholds of up to |4%|. False positive (FP) detections were almost invariant, with respect to sensor FD, being |5%| for sensor deviations of |5%|. Corrections of faulty sensor readings were within SVR model accuracy (AARE < 1%) for SFDI-SVR and |5%| for SFDI-SOM. Although SFDI-SOM has lower detection accuracy, it requires a single overall plant model (or SOM), while providing pictorial representation of plant operation and depiction of faulty operational trajectories.

Pascual X, Gu H, Bartman A, Zhu A, Rahardianto A, Giralt J, Rallo R, Christofides PD, Cohen Y. Fault Detection and Isolation in a Spiral-Wound Reverse Osmosis (RO) Desalination Plant. Ind. Eng. Chem. Res., 2014, 53 (8), 3257–3271

  • DOI: 10.1021/ie403603x

Association Rule Mining of Cellular Responses induced by Metal and Metal Oxide Nanoparticles

2013-10-23 03.50.43 pmRelationships among fourteen different biological responses (including ten signaling pathway activities and four cytotoxicity effects) of murine macrophage (RAW264.7) and bronchial epithelial (BEAS-2B) cells exposed to six metal and metal oxide nanoparticles (NPs) were analyzed using both statistical and data mining approaches. Both the pathway activities and cytotoxicity effects were assessed using high-throughput screening (HTS) over an exposure period up to 24 h and concentration range of 0.39-200 mg/L. HTS data were processed by outlier removal, normalization, and hit-identification (for significantly regulated cellular responses) to arrive at a reliable multiparametric bioactivity profiles for the NPs. Association rules mining was then applied to the bioactivity profiles followed by a pruning process to remove redundant rules. The non-redundant association rules indicated that “significant regulation” of one or more cellular responses imply regulation of other (associated) cellular response types. Pairwise correlation analysis (via Pearson’s χ2 test) and self-organizing map clustering of the different cellular response types indicated consistency with those cellular response types identified in the non-redundant association rules. Furthermore, in order to explore the potential use of association rules as a tool for data-driven hypothesis generation, specific pathway activity experiments were carried out for ZnO NPs. The experimental results confirmed the association rule identified for the p53 pathway and mitochondrial superoxide levels (via MitoSox reagent) and further revealed that blocking of the transcriptional activity of p53 lowered the MitoSox signal. The present approach of using association rule mining for data-driven hypothesis generation has important implications for streamlining multi-parameter HTS assays, improving understanding of NPs toxicity mechanisms, and selection of endpoints for the development of nanomaterial structure-activity relationships.

Liu R, France B, George S, Rallo R, Zhang H, Xia T, Nel A, Bradley K, Cohen Y. Association Rule Mining of Cellular Responses induced by Metal and Metal Oxide Nanoparticles. Analyst, 2013 139 (5), 943-953

  • DOI: 10.1039/C3AN01409F

HDAT: web-based high-throughput screening data analysis tools

2013-10-22 04.14.32 pmThe increasing utilization of high-throughput screening (HTS) in toxicity studies of engineered nano-materials (ENMs) requires tools for rapid and reliable processing and analyses of large HTS datasets. In order to meet this need, a web-based platform for HTS data analyses tools (HDAT) was developed that provides statistical methods suitable for ENM toxicity data. As a publicly available computational nanoinformatics infrastructure, HDAT provides different plate normalization methods, various HTS summarization statistics, self- organizing map (SOM)-based clustering analysis, and visualization of raw and processed data using both heat map and SOM. HDAT has been successfully used in a number of HTS studies of ENM toxicity, thereby enabling analysis of toxicity mechanisms and development of structure–activity relationships for ENM toxicity. The online approach afforded by HDAT should encourage standardization of and future advances in HTS as well as facilitate convenient inter-laboratory comparisons of HTS datasets.

Liu R, Hassan T, Rallo R, Cohen Y. HDAT: web-based high-throughput screening data analysis tools. Computational Science & Discovery. 2013, 6 014006