part1

• Chemical informatics


  The application of information technology to chemistry
• Chemometrics


  The application of statistical methods to chemical data in order to derive predictive models or descriptors
• Chemoinformatics, chemometrics and chemical informatics are related but distinct fields of research
• The term "chemoinformatics" was introduced in 1998 by Brown, who defined it as the combination of "all the information resources that a scientist needs to optimize the properties of a ligand to become a drug"
• Chemoinformatics focuses on decision support by computer and drug discovery relevance, while chemical informatics lacks the specific drug discovery focus
• Spectrum of chemoinformatics


  • Chemical data collection, analysis, and management
  • Data representation and communication
  • Molecular modeling and simulation
  • Structure-activity relationship analysis
  • Virtual screening and compound selection
  • Reaction informatics

  
• As an interdisciplinary field, chemoinformatics involves computational scientists, chemists, and biologists
• Sources that a scientist needs to optimize the properties of a ligand to become a drug
• Chemoinformatics


  The application of information technology to chemistry, with a specific focus on drug discovery
• Chemical informatics


  The application of information technology to chemistry, without a specific drug discovery focus
• It is increasingly difficult to distinguish between chemoinformatics, chemical informatics, and chemometrics, particularly as far as method development is concerned
• Spectrum of chemoinformatics


  • Chemical data collection, analysis, and management
  • Data representation and communication
  • Database design and organization
  • Chemical structure and property prediction (including drug-likeness)
  • Molecular similarity and diversity analysis
  • Compound or library design and optimization
  • Database mining
  • Compound classification and selection
  • Qualitative and quantitative structure-activity or – property relationships
  • Information theory applied to chemical problems
  • Statistical models and descriptors in chemistry
  • Prediction of in vivo compound characteristics

  
• Chemoinformatics includes all concepts and methods designed to interface theoretical and experimental programs involving small molecules
• The evolution of chemoinformatics as an independent discipline will much depend on its ability to demonstrate a measurable impact on experimental chemistry programs, regardless of whether these are in pharmaceutical research or elsewhere
• Hierarchy of bio- and chemoinformatics research

  • DNA sequence
  • Molecular composition
  • Connectivity (graph)
  • Molecular similarity
  • Chemotype
  • Structure
  • Interaction
  • Specific activity
  • Drug
  • Protein sequence
  • Sequence similarity
  • Family
  • Structure
  • Interaction
  • Function
  • Intervention
• Many algorithms and computational techniques used in chemoinformatics are also used for many applications in bioinformatics
• Informatics research and development in the life sciences is expected to become much more global in the future
• Scientific origins of chemoinformatics

  • Quantitative structure-activity relationship (QSAR) analysis
  • Chemical structure storage and retrieval
  • 2D substructure and 3D pharmacophore searching
  • Clustering methods for chemical applications
  • Molecular similarity analysis
  • Molecular diversity and dissimilarity analysis
• Molecular descriptors


  Computational descriptors of molecular structure, physical or chemical properties, or pharmacophores
• Chemical space


  1. dimensional reference space into which molecular data sets are projected for analysis or design

  
• Types of molecular descriptors

  • Physical properties
  • Atom and bond counts
  • Pharmacophore features
  • Charge descriptors
  • Connectivity and shape descriptors

  
• There are no generally preferred descriptor spaces for chemoinformatics applications and it is usually required to generate reference spaces for specific applications on a case-by-case basis
• Similar Property Principle


  Molecules having similar structures and properties should also exhibit similar activity
• Similarity coefficients


  • Tanimoto coefficient
  • Dice coefficient
  • Cosine coefficient
• Molecular similarity, dissimilarity, and diversity

  • Similar molecules can be identified by application of distance functions and analysis of nearest neighbors in chemical space
  • Dissimilar molecules can be identified by maximizing the distance between them in chemical space
  • Molecular diversity refers to the overall spread of a compound collection in chemical space
• Molecular similarity analysis


  The hallmarks of
• Molecular similarity assessment


  1. Descriptor combinations expressed as bit strings (fingerprints)

  2. Test molecule assigned characteristic bit pattern

  3. Pair-wise molecular similarity quantified by overlap of bit strings using similarity metrics (coefficients)
• Similarity metrics (coefficients)


  • Shown in Table 1.4
• ni and nj


  Number of bits set on for molecules i and j, respectively
• nij


  Number of bits in common to both molecules
• Similarity coefficient values


  Range from zero (no overlap; no similarity) to one (complete overlap; identical or very similar molecules)
• The most widely used metric in chemoinformatics is the Tanimoto coefficient
• Molecular similarity


  Identified by application of distance functions and analysis of nearest neighbors in chemical space
• Molecular diversity


  Attempts to either select different compounds from a given population or evenly populate a given chemical space with candidate molecules
• Diversity selection and design

  Using distance functions to select compounds at least a pre-defined minimum distance away from others or maximize average inter-compound distances
• Diversity selection and design

  1. Dividing descriptor axes into evenly spaced value intervals (binning) to produce n-dimensional subsections (cells) of chemical space

  2. Selecting a representative compound from each populated cell or populating cells as evenly as possible with computed molecules
• Molecular diversity is a global concept, while molecular similarity analysis explores pair-wise relationships
• Dissimilarity


  The inverse of molecular similarity, addressing which molecule in a collection is most dissimilar from a given compound or set of compounds
• Dissimilarity-based compound selection


  1. Initially selecting a seed compound, then calculating dissimilarity between the seed and all others and selecting the most dissimilar one

  2. Repeating the process to obtain a subset of desired size
• High-dimensional chemistry spaces might often be too complex for carrying out meaningful and interpretable analyses