Structure Activity Relationship Analysis and Development

The structure of a chemical implicitly determines its physical and chemical properties and reactivities. Structure-activity relationships (SARs) and quantitative structure–activity relationships (QSARs) are basic and theoretical models to drug research and development, which can be used to predict the physicochemical, biological, and other properties of chemicals, and guide compound optimization. The former one is an (qualitative) association between a chemical substructure and the potential of a chemical. The latter one is a mathematical model that quantitatively relates a quantitative numerical measure of chemical structure, which has been commonly used in the absence of available data for prioritization, classification, and assessment. Classical QSARs methods are commonly used to build linear SAR models for compound series and prioritize analogs.

SARs has long been used to design chemicals with commercially desirable properties with expected pharmacologic and therapeutic activities. Moreover, computer-based modeling methods, as well as AI-driven approaches relating chemical structure to qualitative biological activity and quantitative biological potency have been widely applied in diverse problem settings. CD ComputaBio provides intelligent services of structure activity relationship analysis and development for our clients in different stage of drug discovery, especially for screening large numbers of chemicals and decision making during chemical optimization.

Goals

Identify the “key” compounds.
Propose possible “hole” of the chemical scaffold.
Delineate classes of active chemicals representing distinct biological and chemical mechanism domains.
Determine the structural features and properties responsible for modulating activity.
Modeling approaches that are tailored to issues in drug research and development.

Figure 1 Communication and Modeling Platform for Medicinal Chemists

Application Scenarios

Quick structure clustering
Similarity/substructure search
Properties predictions
Different visualization

Model calculation
Protein binding site analysis
Hit selection
Molecule docking and scoring

Structure Activity Relationship Analysis

A variety of computational methods are available to aid in SAR analysis and compound design. Different statistical and modeling approaches have been introduced to monitor SAR progression of evolving compound series. AI-powered methods are also increasingly applied to model nonlinear SARs and predict novel active compounds.

SAR Analysis Endpoints

Enzymatic assay
Cellular assay
Selectivity
Solubility
Permeability

H(M/R)LM
TDI
DDI
hERG binding
Off target panel

SAR Analysis Methods

R-group decomposition
Structure clustering
Matched molecular pairs
Activity cliffs
Similarity matrix
Scaffold tree
Automatic core scaffold determination

Structure activity landscape index
Molecular grid maps
Principle component analysis
SAR visualization
Machine learning algorithms
Numerical SAR analysis

SAR Analysis Tools

ICM Molsoft (FOCUS)
Tibco Spotfire
StarDrop
Instant Jchem
Schrodinger
MOE

Cresset
KNIME
Data Warrior
OSIRIS Property Explorer
MultiCASE (MCASE)
Various internal developed informatics tools & integrations

Structure Activity Relationship Development

The process of developing a SAR is one of attempting to understand and reveal how properties relevant to activity are encoded within and determined by the chemical structure. New SAR modules are constantly being developed. These modules should be tested and refined for use in further validation experiments. Various methodologies have been developed to systematically extract structurally related compound series from data sets of any composition, visualize SAR patterns, and generate virtual candidate compounds. Experts from CD ComputaBio are dedicated to making it possible to comprehensively organize and explore large data sets, visualize SARs, and select candidate compounds for practical applications.