A novel variable reduction method adapted from space-filling designs

Ballabio, Davide; Consonni, Viviana; Mauri, Andrea; Claeys‐Bruno, Magalie; Sergent, Michelle; Todeschini, Roberto

doi:10.1016/j.chemolab.2014.05.010

Cited by 77 publications

(55 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Descriptor matrix was refined by eliminating descriptors having correlation lower than 0.3 leaving 4248 descriptors. V-WSP variable reduction MATLAB routine which is an unsupervised variable reduction based on V-WSP algorithm (Ballabio et al, 2014) was subsequently applied on 4248 descriptors with 0.85 absolute correlations which gave 357 most suitable descriptors. The dataset was split into training set of 44 compounds and test set of 13 compounds.…”

Section: Variable Selection and Model Developmentmentioning

confidence: 99%

4D-QSAR studies of coumarin derivatives as HIV-1 integrase 3′-processing inhibitors

Patil

Sawant

2015

Med Chem Res

View full text Add to dashboard Cite

In a pursuit of combating acquired immunodeficiency syndrome, the targets human immunodeficiency virus (HIV) protease, reverse transcriptase and HIV specific proteins were exhaustively explored. Resistance through mutations in these targets has impeded the further developments in HIV research. HIV-1 integrase (HIV-1 IN) has emerged as a newer target and 3 0 -processing, and strand transfer inhibition has proved potential intervention in HIV replication. In present study, in line with our interests in coumarin derivatives as potential bioactive molecules, 4D-quantitative structure-activity relationship (4D-QSAR) is proposed. Fifty-seven structurally diverse coumarin derivatives were subjected to 4D-QSAR studies. Quantum mechanics-based geometry optimization and molecular dynamics simulation were carried out on individual compound. The conformational ensemble generated for each compound was aligned with most active compound, and Coulombic and Lennard-Jones interaction energy descriptors were computed. After selecting the best variables in MATLAB, partial least square regression (PLS) analysis was carried out on 44 training set and 13 test set compounds. The model with ten latent variables was found best with R 2 calculated = 0.903015, R 2 crossvalidated = 0.599553, R 2 predicted = 0.688525, rootmean-square error (RMSE) calculated = 0.21276, RMSE predicted = 0.371579 and prediction bias = -0.15362. Docking studies were carried out on AutoDock Vina, which were in good agreement with the PLS model, suggesting the importance of few descriptors of Coulombic interaction energy and VWD interactions with VAL79, VAL77, ARG199 and GLU157. These results may be useful in designing more potent HIV-1 IN inhibitors.

show abstract

Section: Variable Selection and Model Developmentmentioning

confidence: 99%

4D-QSAR studies of coumarin derivatives as HIV-1 integrase 3′-processing inhibitors

Patil

Sawant

2015

Med Chem Res

View full text Add to dashboard Cite

show abstract

“…Following descriptor generation, the datasets were curated by removing variables with null values as well as zero variance variables, i.e., descriptors with a standard deviation <0.0001. This is referred to as the “Original dataset.” The original datasets were then subjected to variable—Wootton, Sergent, Phan‐Tan‐Luu algorithm (V‐WSP) reduction, an unsupervised variable reduction method that allows for the elimination of variables based on multicollinearity . This was performed via a grid search to find the correlation coefficient threshold for the descriptors for which the Procrustes index was lower than 0.2.…”

Section: Methodsmentioning

confidence: 99%

“…The Procrustes index is a statistical measure that allows for the assessment of the degree of comparability between the original and reduced datasets based on informational content. A Procrustes value of 1 indicates complete dissimilarity and 0 indicates that both datasets are identical . The grid search was carried out separately for the descriptors of different domains to avoid removing chance correlations due to the presence of similar amino acids in different domains of the variable regions, i.e., VH and VL domains.…”

Section: Methodsmentioning

confidence: 99%

Cross‐Interaction Chromatography‐Based QSAR Model for Early‐Stage Screening to Facilitate Enhanced Developability of Monoclonal Antibody Therapeutics

Kizhedath

Karlberg

2019

Biotechnology Journal

View full text Add to dashboard Cite

Monoclonal antibodies (mAbs) constitute a rapidly growing biopharmaceutical sector. However, their growth is impeded by developability issues such as polyspecificity and lack of solubility, which leads to attrition as well as manufacturing failures. In this study a multitool hybrid quantitative structure–activity relationship (QSAR) model development framework is described. This framework uses four novel datasets derived from the primary sequences of IgG1‐κ‐humanized mAbs with varying degrees of resolutions. Unsupervised pattern recognition is first performed on the descriptor sets to visualize any intrinsic property‐based clustering, followed by regression of descriptors against cross‐interaction chromatography (CIC) retention times. Model optimization is performed via unsupervised variable reduction followed by supervised variable selection. Finally, the models and datasets are benchmarked based on the regression model performance metrics such as R2, Q2, and RMSE. The results show that datasets containing localized descriptors rather than averaged value over the entire protein have better predictive performance of CIC retention behavior with R2 > 0.8 and RMSE < 0.3. Furthermore, the results indicate the physicochemical, electronic, and topological properties of hypervariable regions of antibodies that contribute most to the CIC retention times. The results of these studies could contribute to early‐stage screening and better design of mAbs.

show abstract

“…al. [29]. After running the V-WSPtool successfully, a log file is generated that consists of names of all the descriptorswhich are removed based on variance cut-off (i.e.…”

Section: 4data Pre-treatment -V-wsp Toolmentioning

confidence: 99%

“…In this step, variance is calculated for each descriptor j(where, …,m)and the descriptors with variance less than user specified variance cut-off value are removed.Further, the V-WSPalgorithmis applied on the modified data matrix with n rows and p columns (p= number of descriptors remaining after removing constant descriptors), to remove intercorrelated descriptors. The V-WSP algorithm mentioned below is according to the literature [29]. 1.…”

mentioning

confidence: 99%

“NanoBRIDGES” software: Open access tools to perform QSAR and nano-QSAR modeling

Ambure

Aher

Gajewicz

et al. 2015

Chemometrics and Intelligent Laboratory Systems

152

View full text Add to dashboard Cite

A novel variable reduction method adapted from space-filling designs

Cited by 77 publications

References 29 publications

4D-QSAR studies of coumarin derivatives as HIV-1 integrase 3′-processing inhibitors

4D-QSAR studies of coumarin derivatives as HIV-1 integrase 3′-processing inhibitors

Cross‐Interaction Chromatography‐Based QSAR Model for Early‐Stage Screening to Facilitate Enhanced Developability of Monoclonal Antibody Therapeutics

“NanoBRIDGES” software: Open access tools to perform QSAR and nano-QSAR modeling

Contact Info

Product

Resources

About