Abstract
In this work we report the first attempt to study the effect of activity cliffs over the generalization ability of machine learning (ML) based QSAR classifiers, using as study case a previously reported diverse and noisy dataset focused on drug induced liver injury (DILI) and more than 40 ML classification algorithms. Here, the hypothesis of structure-activity relationship (SAR) continuity restoration by activity cliffs removal is tested as a potential solution to overcome such limitation. Previously, a parallelism was established between activity cliffs generators (ACGs) and instances that should be misclassified (ISMs), a related concept from the field of machine learning. Based on this concept we comparatively studied the classification performance of multiple machine learning classifiers as well as the consensus classifier derived from predictive classifiers obtained from training sets including or excluding ACGs. The influence of the removal of ACGs from the training set over the virtual screening performance was also studied for the respective consensus classifiers algorithms. In general terms, the removal of the ACGs from the training process slightly decreased the overall accuracy of the ML classifiers and multi-classifiers, improving their sensitivity (the weakest feature of ML classifiers trained with ACGs) but decreasing their specificity. Although these results do not support a positive effect of the removal of ACGs over the classification performance of ML classifiers, the “balancing effect” of ACG removal demonstrated to positively influence the virtual screening performance of multi-classifiers based on valid base ML classifiers. Specially, the early recognition ability was significantly favored after ACGs removal. The results presented and discussed in this work represent the first step towards the application of a remedial solution to the activity cliffs problem in QSAR studies.
Keywords: Activity cliffs, activity cliffs generators, chemoinformatics, machine learning, QSAR, virtual screening.
Current Pharmaceutical Design
Title:Probing the Hypothesis of SAR Continuity Restoration by the Removal of Activity Cliffs Generators in QSAR
Volume: 22 Issue: 33
Author(s): Maykel Cruz-Monteagudo, José L. Medina-Franco, Yunier Perera-Sardiña, Fernanda Borges, Eduardo Tejera, Cesar Paz-y-Miño, Yunierkis Pérez-Castillo, Aminael Sánchez-Rodríguez, Zuleidys Contreras-Posada and M. Natália D. S. Cordeiro
Affiliation:
Keywords: Activity cliffs, activity cliffs generators, chemoinformatics, machine learning, QSAR, virtual screening.
Abstract: In this work we report the first attempt to study the effect of activity cliffs over the generalization ability of machine learning (ML) based QSAR classifiers, using as study case a previously reported diverse and noisy dataset focused on drug induced liver injury (DILI) and more than 40 ML classification algorithms. Here, the hypothesis of structure-activity relationship (SAR) continuity restoration by activity cliffs removal is tested as a potential solution to overcome such limitation. Previously, a parallelism was established between activity cliffs generators (ACGs) and instances that should be misclassified (ISMs), a related concept from the field of machine learning. Based on this concept we comparatively studied the classification performance of multiple machine learning classifiers as well as the consensus classifier derived from predictive classifiers obtained from training sets including or excluding ACGs. The influence of the removal of ACGs from the training set over the virtual screening performance was also studied for the respective consensus classifiers algorithms. In general terms, the removal of the ACGs from the training process slightly decreased the overall accuracy of the ML classifiers and multi-classifiers, improving their sensitivity (the weakest feature of ML classifiers trained with ACGs) but decreasing their specificity. Although these results do not support a positive effect of the removal of ACGs over the classification performance of ML classifiers, the “balancing effect” of ACG removal demonstrated to positively influence the virtual screening performance of multi-classifiers based on valid base ML classifiers. Specially, the early recognition ability was significantly favored after ACGs removal. The results presented and discussed in this work represent the first step towards the application of a remedial solution to the activity cliffs problem in QSAR studies.
Export Options
About this article
Cite this article as:
Cruz-Monteagudo Maykel, Medina-Franco L. José, Perera-Sardiña Yunier, Borges Fernanda, Tejera Eduardo, Paz-y-Miño Cesar, Pérez-Castillo Yunierkis, Sánchez-Rodríguez Aminael, Contreras-Posada Zuleidys and Cordeiro D. S. M. Natália, Probing the Hypothesis of SAR Continuity Restoration by the Removal of Activity Cliffs Generators in QSAR, Current Pharmaceutical Design 2016; 22 (33) . https://dx.doi.org/10.2174/1381612822666160509124337
DOI https://dx.doi.org/10.2174/1381612822666160509124337 |
Print ISSN 1381-6128 |
Publisher Name Bentham Science Publisher |
Online ISSN 1873-4286 |
Call for Papers in Thematic Issues
"Tuberculosis Prevention, Diagnosis and Drug Discovery"
The Nobel Prize-winning discoveries of Mycobacterium tuberculosis and streptomycin have enabled an appropriate diagnosis and an effective treatment of tuberculosis (TB). Since then, many newer diagnosis methods and drugs have been saving millions of lives. Despite advances in the past, TB is still a leading cause of infectious disease mortality ...read more
Current Pharmaceutical challenges in the treatment and diagnosis of neurological dysfunctions
Neurological dysfunctions (MND, ALS, MS, PD, AD, HD, ALS, Autism, OCD etc..) present significant challenges in both diagnosis and treatment, often necessitating innovative approaches and therapeutic interventions. This thematic issue aims to explore the current pharmaceutical landscape surrounding neurological disorders, shedding light on the challenges faced by researchers, clinicians, and ...read more
Emerging and re-emerging diseases
Faced with a possible endemic situation of COVID-19, the world has experienced two important phenomena, the emergence of new infectious diseases and/or the resurgence of previously eradicated infectious diseases. Furthermore, the geographic distribution of such diseases has also undergone changes. This context, in turn, may have a strong relationship with ...read more
Melanoma and Non-Melanoma Skin Cancer Treatment: Standard of Care and Recent Advances
In this thematic issue, we aim to provide a standard of care of the diagnosis and treatment of melanoma and non-melanoma skin cancer. The editor will invite authors from different countries who will write review articles of melanoma and non-melanoma skin cancers. The Diagnosis, Staging, Surgical Treatment, Non-Surgical Treatment all ...read more
- Author Guidelines
- Graphical Abstracts
- Fabricating and Stating False Information
- Research Misconduct
- Post Publication Discussions and Corrections
- Publishing Ethics and Rectitude
- Increase Visibility of Your Article
- Archiving Policies
- Peer Review Workflow
- Order Your Article Before Print
- Promote Your Article
- Manuscript Transfer Facility
- Editorial Policies
- Allegations from Whistleblowers
- Announcements
Related Articles
-
Pomegranate Use to Attenuate Bone Loss in Major Musculoskeletal Diseases: An Evidence-Based Review
Current Drug Targets N-Myristoyltransferase: A Novel Target
Mini-Reviews in Medicinal Chemistry The Antibacterial Potential of Ciprofloxacin Hybrids against <i>Staphylococcus aureus</i>
Current Topics in Medicinal Chemistry Phytochemistry and Pharmacology of Indian Traditional Plant Hyssop (<i>Hyssopus officinalis</i> L.): A Review
The Natural Products Journal Demographic, Clinical, and Investigational Characteristics of COVID-19- related Guillain-Barré Syndrome with Differences from Typical and Another Virus-related Guillain-Barré Syndrome
Infectious Disorders - Drug Targets Characteristic Peptides of Protein Secondary Structural Motifs
Protein & Peptide Letters What we have Learned from Animal Models of Mycoplasma pneumoniae Disease: Virulence Mechanisms and Host Responses
Current Pediatric Reviews Design, Synthesis and Molecular Docking Studies of 2-Aryl/Heteroaryl-Ethyl 6-Chloroquinoline-4-Carboxylates as Potential Antimalarial Agents
Letters in Drug Design & Discovery Interactions between ACE2 and SARS-CoV-2 S Protein: Peptide Inhibitors for Potential Drug Developments Against COVID-19
Current Protein & Peptide Science Chemical Informatics and the Drug Discovery Knowledge Pyramid
Combinatorial Chemistry & High Throughput Screening Vaccine Development Against SARS-CoV-2: From Virology to Vaccine Clinical Trials
Coronaviruses An Overview of Recent Dipeptidyl Peptidase-IV Inhibitors: Linking their Structure and Physico-Chemical Properties with SAR, Pharmacokinetics and Toxicity
Current Topics in Medicinal Chemistry Compound Activity Prediction Using Models of Binding Pockets or Ligand Properties in 3D
Current Topics in Medicinal Chemistry Measurement of Extravascular Lung Water in Critically Ill Patients
Current Respiratory Medicine Reviews Phenothiazines and Structurally Related Compounds as Modulators of Cancer Multidrug Resistance
Current Drug Targets Nanotechnology for Delivery of Drugs and Biomedical Applications
Current Clinical Pharmacology Chronic Inflammatory Diseases and the Acute Respiratory Distress Syndrome (ARDS)
Current Pharmaceutical Design Management of Food-Induced Anaphylaxis: Unsolved Challenges
Current Clinical Pharmacology An Insight into the Dermatological Applications of Neem: A Review on Traditional and Modern Aspect
Recent Advances in Anti-Infective Drug Discovery Advancements in COVID-19 Testing: An In-depth Overview
Current Pharmaceutical Biotechnology