For decades, the drug development process has usually been like a battle between trial and error. Only a very few of the millions of drug candidates can be successfully developed. Recently, researchers have used computer technology to make a good start for drug research and development. By analyzing the chemical structure of a drug, researchers can understand whether it may bind or "docking" a biological target (such as a protein). Such an algorithm is particularly useful to discover potential toxic and side effects caused by accidental binding of drugs to structurally similar non-target proteins.
Recently, researchers have proposed a computer simulation work plan to evaluate the connection between billions of possible drugs and proteins based on drug information and protein information stored in public databases. Timothy Cardozo, a pharmacologist at the Langone Medical Center at New York University, attended the National Institutes of Health in Bethesda, Maryland on November 19 "High Risk-High Reward Symposium" seminar, and showed this computer simulation technology at the meeting; he pointed out that this is the largest computer simulation work done by humans in history. This work eventually created a website called Drugable (drugable.com), which was supported by the National Library of Medicine (NLM); although Drugable is still in beta, it will eventually Open to the public, so that researchers can predict how and where they work in the body based on the chemical structure of the compound (see "Drug Information Mining").
Drug Information Mining
The researchers used Google's supercomputer to detect billions of drug-protein interactions.
Cardozo admits that computer simulation is only an initial step in the new drug development process. After predicting whether a certain protein can bind to the compound, drug developers must test the effect of the drug on this protein in the cell, so as to observe the actual effect of the drug on the protein function, and the dosage of the drug in different situations. Wait. Animal experiments are then carried out; if you are lucky, the researchers can then conduct human experiments. But Brian Shoichet, a computational biologist at the University of California, San Francisco, points out that pharmaceutical companies often hold patents on these additional data. He pointed out that although some public databases (such as PubChem maintained by NLM) hold automated detection results of the interaction between drugs and yeast cell proteins, they contain some inaccurate results and false positive results.
Nevertheless, scientists have confirmed that computer simulation methods can provide some shortcuts for drug development. In 2012, Shoichet collaborated with researchers at the Novartis Institutes for BioMedical Research in Cambridge, Massachusetts to jointly develop an algorithm that can predict the side effects of drugs based on the similarity of the chemical structure between drugs . When the researchers tested the interaction between 656 approved drugs and 73 biological targets, they found that the algorithm can predict hundreds of previously unknown interactions, and about half of the predictions are It is confirmed that these side effects are true (E. Lounkine et al. Nature 486. 361-367; 2012). Shoichet also pointed out that for known drugs, this type of calculation provides a quick way to confirm drug-protein interactions for further research.
Drugable attempts to predict how untested compounds interact with proteins in the body, and such predictions are more challenging. When establishing the Drugable website, the Cardozo research team selected approximately 600,000 compound molecules from the PubChem database and the ChEMBL database of the European Bioinformatics Institute (which contains information on millions of published compounds) And evaluated the ability of these molecules to bind to the 7000 structural "pockets" of human proteins in the database. Computing giant Google has provided researchers with the equivalent of 100 million hours of supercomputer processor time to help them complete this huge job.
The research team proposed to use docking scores to rank the strength of more than 4 billion potential drug-protein interactions. Subsequently, the research team cross-referenced the target protein that it predicted with the target protein in the Gene Expression Omnibus database to indicate the expression site of different protein-coding genes in the body. Cardozo pointed out that this allows researchers to predict the possible parts of the drug: if Drugable finds that there is an interaction between a certain drug and a highly expressed protein in a certain part, then this drug is likely to Organizations play a role.
Researcher Jeremy Jenkins of the Novartis Institute pointed out that pharmaceutical companies have been carrying out similar computer predictions for many years. But he also pointed out: Although Novartis has a library of 1.5 million public compounds and patented compounds, it has never analyzed so many proteins and drugs at once like Drugable.
Cardozo hopes that Drugable can provide special assistance in the evaluation of antipsychotic drugs, because these drugs usually work in ways that are difficult to measure. As a demonstration, the Cardozo research team used the Drugable algorithm to predict target interactions for two drugs commonly used to treat schizophrenia, clozapine and chlorpromazine.
As expected, the Drugable prediction results indicate that the two drugs have the strongest binding ability to the neurotransmitter dopamine (serotonin) and dopamine (dopamine) receptors, which are usually expressed in advanced information Processing in the brain area. However, Drugable found that clozapine, which can also be used to treat mood disorders (such as depression), can strongly bind to a dopamine receptor called DRD4, and this special receptor is mainly expressed in the brain's emotion regulation Area-in the pineal gland (pineal gland).
The research team also found that clozapine can bind to receptors in the brain that regulate saliva secretion; excessive saliva secretion is a known side effect of clozapine. Although researchers have previously explained the effects of clozapine on emotion regulation and saliva secretion from a biochemical point of view, Cardozo believes that Drugable can be used to reveal the most reasonable mechanism of drug action.
Banana Holders,Banana Holder,Plastic Banana Holder,Banana Rack Stand
Jiangmen Jianghai Jianshang Houseware Co.,LTD. , https://www.jm-jianshang.com