<h5 id="101-protein-information">10.1 Protein Information</h5> <ul> <li>Protein is a complex molecule that plays a critical role in the body.</li> <li>It is composed of amino acids, which are linked together by peptide bonds to form a polypeptide chain.</li> <li>The sequence of these amino acids determines the protein’s structure and function.</li> <li>Proteins can be classified into three main groups: fibrous, globular, and membrane.</li> <li>The shape of a protein is crucial to its function, and this shape is maintained by various interactions, including disulfide bonds, hydrogen bonds, and hydrophobic interactions.</li> </ul> <p>======</p> <h5 id="1011-introduction">10.1.1 Introduction</h5> <ul> <li>Protein informatics involves gathering information about proteins using methods of information technology.</li> <li>It has significantly aided in identifying the geometrical location of functional sites, biochemical and biological functions of hypothetical proteins.</li> <li>Protein informatics has enabled the determination of tertiary structures of many hypothetical proteins, which were previously unclear.</li> <li>The development of protein informatics has been supported by heterogeneous databases and various descriptors of amino acid sequences, tertiary structures, and pathways on the proteome scale.</li> <li>Protein informatics has been instrumental in understanding molecular functions of proteins, which were not possible with conventional methods.</li> </ul> <p>======</p> <h5 id="1012-protein-data-types">10.1.2 Protein data types</h5> <ul> <li>Protein data can be of various types:</li> </ul> <p>======</p> <h5 id="1013-computational-prediction-of-protein-structures">10.1.3 Computational prediction of protein structures</h5> <ul> <li>Protein structure prediction is a key focus in bioinformatics, aiming to determine how amino acid sequences specify protein structure and function.</li> <li>This prediction can be done using bioinformatics tools, even when only the gene sequence is known.</li> <li>Numerous computational tools are available for predicting structural and physicochemical properties of proteins.</li> <li>The major advantages of computational methods include their quick turnaround time, cost-effectiveness, and suitability for high throughput screening.</li> <li>These methods are particularly useful for predicting the structure of hypothetical proteins.</li> </ul> <p>======</p> <h5 id="10131-primary-structure-prediction">10.1.3.1 Primary structure prediction</h5> <ul> <li>Isoelectric point (pI) is the pH at which a protein has no net charge and is stable and compact. A pI < 7 indicates an acidic protein, while a pI > 7 indicates a basic protein.</li> <li>The aliphatic index (AI) is a measure of the relative volume of a protein occupied by aliphatic side chains, and is a positive factor for thermal stability. A high AI indicates that a protein may be stable over a wide temperature range.</li> <li>The instability index is a measure of protein stability in a test tube, with unstable proteins having a higher occurrence of certain dipeptides. A protein with an instability index smaller than 40 is predicted to be stable, while a value above 40 indicates potential instability.</li> <li>The Grand Average Hydropathy (GRAVY) value is a measure of a protein’s hydrophobicity, calculated as the sum of hydropathy values of all amino acids divided by the number of residues. A low GRAVY value indicates a greater potential for interaction with water.</li> <li>These properties can be calculated using the ProtParam tool of the ExPASy Proteomics Server for protein primary structure prediction.</li> </ul> <p>======</p> <h5 id="10132-secondary-structure-prediction">10.1.3.2 Secondary Structure Prediction</h5> <ul> <li>Protein secondary structure prediction is important for revealing the functions of proteins with unknown structures.</li> <li>It is also a step towards protein 3-dimensional structure prediction.</li> <li>Some common protein secondary structure prediction tools are APSSP, CFSSP, SOPMA, and GOR.</li> <li>These tools predict the secondary structure of proteins based on their amino acid sequence.</li> <li>The prediction of protein secondary structure can help in understanding the protein’s behavior and functionality.</li> </ul> <p>======</p> <h5 id="10133-three-dimensional-3d-structure-prediction">10.1.3.3 Three dimensional (3D) Structure Prediction</h5> <ul> <li> <p>Homology modelling is a method used to predict protein 3D structure based on amino acid sequence similarity with known structures. It is advantageous as it does not require knowledge of physical determinants.</p> </li> <li> <p>Fold prediction, also known as fold recognition, aligns structures and uses ’threading’ to force the sequence of a protein with unknown structure to take the conformation of a known structure’s backbone. This method is more compute-intensive but provides confidence in physical viability.</p> </li> <li> <p>De novo protein structure prediction is an algorithmic process that predicts tertiary structure from the primary sequence. QUARK is a tool for this method, constructing models from small fragments using replica-exchange Monte Carlo simulation.</p> </li> <li> <p>Protein structure is recorded in PDB files, which store 3D coordinates from X-ray crystallography, NMR, and theoretical models. The PDB database is linked with protein databases for homology modelling and threading.</p> </li> <li> <p>Domain prediction, using tools like InterPRO scan and CDD search, identifies distinct functional and structural units in a protein, providing valuable information for structure, function, evolution, and design prediction.</p> </li> </ul> <p>======</p> <h5 id="102-cheminformatics">10.2 Cheminformatics</h5> <ul> <li>Cheminformatics is a field that combines chemistry and information science.</li> <li>It involves the use of various computational tools and methods to analyze and manage chemical data.</li> <li>The primary goal is to extract useful information from complex chemical data and represent it in a more understandable and useful format.</li> <li>This field uses various algorithms, databases, and visualization tools to store, search, and analyze chemical data.</li> <li>Equations and formulae are not explicitly provided in the text, but cheminformatics often involves the application of mathematical models to chemical data.</li> </ul> <p>======</p> <h5 id="1021-introduction">10.2.1 Introduction</h5> <ul> <li>Cheminformatics is a interdisciplinary field that combines principles of physics, chemistry, biology, mathematics, biochemistry, statistics and informatics.</li> <li>It involves the use of computational and informational techniques to understand problems of chemistry.</li> <li>Cheminformatics is useful in drug discovery where large numbers of compounds are evaluated for interaction with target cellular molecules.</li> <li>It deals with information on physical properties, three-dimensional molecular and crystal structures, chemical reaction pathways, etc.</li> <li>Virtual screening, a part of cheminformatics, identifies and evaluates the best candidates for a particular property or reaction from large libraries of real and virtual molecules.</li> </ul> <p>======</p> <h5 id="1022-storing-and-managing-the-chemical-data">10.2.2 Storing and managing the chemical data</h5> <ul> <li>PubChem is a database of chemical molecules that maintains information on substances, compounds, and BioAssays.</li> <li>ZINC database contains 21 million compounds with features like molecular weight and log P.</li> <li>ChEMBL provides information on 1 million bioactive compounds with 8200 drug targets.</li> <li>NCI database has over $2,75,000$ small molecule structures, useful for cancer/AIDS research.</li> <li>ChemDB contains 5 million chemicals with predicted or experimentally determined physicochemical properties.</li> <li>ChemSpider has 28 million unique chemical entities from over 400 data sources.</li> <li>BindingDB is a binding affinity database of small molecules with 9,108,36 binding data for 6,263 protein targets and 378,980 small molecules.</li> <li>DrugBank contains 6712 drug entries, including 1448 FDA-approved small molecule drugs and 131 FDA-approved biotech drugs.</li> <li>PharmaGKB is a pharmacogenomics knowledge resource that encompasses clinical information of drug molecules.</li> <li>SuperDrug contains approximately 2500 3D-structures of active ingredients of essential marketed drugs.</li> </ul> <p>======</p> <h5 id="1023-why-do-we-need-cheminformatics">10.2.3 Why do we need cheminformatics?</h5> <ul> <li>Cheminformatics is essential for navigating large chemical databases, which contain hundreds of millions of compounds and associated properties.</li> <li>It helps discover patterns within this vast body of literature and data.</li> <li>Pharmaceutical companies utilize cheminformatics for in silico design of new drugs, followed by synthesis and testing.</li> <li>The chemical manufacturing industry benefits from cheminformatics in designing new properties, predicting efficacy, and assessing toxicity of chemicals prior to market release.</li> <li>Cheminformatics tools are crucial for browsing, finding, and identifying the right chemical compounds that meet specific requirements.</li> </ul> <p>======</p> <h5 id="1024-how-to-store-information-on-chemical-compounds">10.2.4 How to store information on chemical compounds?</h5> <ul> <li>Chemical compounds can be drawn on paper or digitally using predefined templates.</li> <li>At the research level, chemical structures are stored as molecular graphs, which are collections of nodes (chemical substances) and edges (movement of information between nodes).</li> <li>These graphs can have subgraphs and may contain cycles or rings. Alternatively, they can be represented as trees with root, branch, and leaf nodes.</li> <li>The graph is then communicated to the computer using a connection table, which includes atomic numbers, bonds, hybridization state, and coordinates.</li> <li>Hydrogen atoms may be omitted in the connection table, making it hydrogensuppressed. Chemical structures can also be represented by linear notation such as SMILES, which uses alphanumeric characters to store information for computation.</li> </ul> <p>======</p> <h5 id="1025-searching-the-structures">10.2.5 Searching the structures</h5> <ul> <li>Commercially available databases, including Cheminformatics, often originate from academic research projects.</li> <li>The simplest task in searching these structures involves extracting information on chemical properties, such as finding physical and chemical properties of a substance or listing all chemical substances within a certain boiling point range.</li> <li>A more complex search involves substructure retrieval, showing all chemical compounds that correspond to a certain functional group, like a methyl group, benzene ring, or alkene backbone.</li> <li>Subgraph isomorphism is a two-stage search process where a general screen eliminates molecules that do not possibly match the substructure query, and a more elaborate subgraph isomerism process finds molecules that truly match a given substructure.</li> <li>Molecule screens are implemented using binary strings of 0s and 1s, called bitstrings.</li> </ul> <p>======</p> <h5 id="1026-searching-the-reactions">10.2.6 Searching the reactions</h5> <ul> <li>Chemists can use reaction databases to search for products and reaction conditions of a given compound.</li> <li>The search can be refined by integrating multiple queries, such as finding all reactions that utilize a specific compound and operate within a certain temperature range.</li> <li>A significant feature of reaction search is atom mapping, which establishes an exact correspondence between reactant atoms and resultant products.</li> <li>Existing cheminformatics tools and databases can retrieve reactions where a certain substructure is converted into products.</li> <li>Information such as solvents, pH, temperature, and pressure can also be searched in reaction queries.</li> </ul> <p>======</p> <h5 id="1027-pharmacophore">10.2.7 Pharmacophore</h5> <ul> <li>A pharmacophore is a conceptual framework that defines molecular features necessary for a ligand’s biological recognition.</li> <li>It is an ensemble of steric and electronic features that interact with a specific biological target, triggering a biological response.</li> <li>Pharmacophore models explain how structurally diverse ligands can bind to a single receptor molecule.</li> <li>A 3D pharmacophore includes features related to spatial orientation, such as charged groups, rings, and hydrophobic regions.</li> <li>It’s crucial to note that a pharmacophore is not a physical molecule or a group of molecules, but a theoretical concept used to describe a therapeutic molecule’s necessary features for interaction with a target.</li> </ul> <p>======</p> <h5 id="1028-lipinskis-rule-of-five-r05">10.2.8 Lipinski’s rule of five (R05)</h5> <ul> <li>The Lipinski’s rule of five (R05) is a set of criteria proposed by Christopher A. Lipinski in 1997 to describe key molecular properties of compounds.</li> <li>R05 provides indicative information about Absorption (A), Distribution (D), Metabolism (M), Excretion (E) and Druglikeness properties of small molecules.</li> <li>The rule includes: 1) Not more than 5 hydrogen bond donors, 2) Not more than 10 hydrogen bond acceptors, 3) Molecular weight below 500 Daltons, 4) Octanol water partition coefficient (log P) of less than 5.</li> <li>A compound is not favoured further if its RO5 score is greater than 1, indicating expected unfavourable performance during absorption, distribution, metabolism and excretion.</li> <li>R05 applies only to finding a chemical compound that has potential to be a successful oral drug, and does not consider pharmacological effect of a drug-like molecule or drugs given through intramuscular and intravenous routes.</li> </ul> <p>======</p> <h5 id="1029-the-journey-of-a-drug">10.2.9 The journey of a drug</h5> <ul> <li>The drug discovery and development process is a long, expensive, and risky journey from the lab to the market.</li> <li>Virtual screening is an in-silico approach used to determine useful compounds for specific purposes, such as drug discovery and industrial applications, by scoring, ranking, and extracting structures through computational methods.</li> <li>Virtual screening consists of a series of filters that eliminate undesirable compounds, starting with a broad set of parameters and moving towards a narrow set to identify a small group of molecules with desired properties.</li> <li>The process may involve general filters to identify drug-like compounds with specific ADME properties, ligand-based methods like machine learning techniques and pharmacophore-based search, and structure-based methods like protein-ligand docking.</li> <li>The discovery of saccharin, an artificial sweetener, was an accidental finding in 1879 by a Russian chemist, Dr. Constantin Fahlberg, who tasted a sweet substance on his hands after working with coal tar. He ran back to the lab and found an impure solution of saccharin, which led to the start of a company and global fame.</li> </ul> <p>======</p>