Automated Drug Repurposing Based On Omics ' Data Mining

This paper hopes to share insight into the steps that are taken by companies, and the strenuous process behind developing an effective new drug.

There are multiple health concerns worldwide and more and more drugs are needed every day. Many drugs however, are extremely expensive to develop, test, and produce. According to the Tufts Center for the Study of Drug Development (2002), it costs up to $802 million to bring a new drug to the market. In 2002, pharmaceutical companies spent $34 billion in research and development (Center-Watch, 2003). In addition to the costs, the overall time from the discovery to approve and market the drug can take up to 15 years.

Drug forecasting is a very important and essential process for any drug for multiple reasons:

With the huge diversity and changeability of human biology, it is impossible to imagine a reality without some mutations, changes, or issues in the organs and tissues of humans. Thus, it rightly follows that medications and pharmaceuticals have been created in an effort to counteract the various ailments and illnesses that people can experience. However, as time has gone on and these pharmaceuticals have become more and more high-tech, regulated, and trusted, they have also become incredibly commercialized. Worse still, medications have become incredibly expensive and can be unattainable for some people.

High throughput screening (HTS) is essential in the process of finding a potential drug candidate, and automation is key for a good screen both in terms of robustness and reproducibility.

Ng, Rick. Drugs: From Discovery to Approval. 3rd ed. Hoboken, NJ: John Wiley & Sons, 2015.

- A computer can display the molecular structure of any drug from a listing of thousands contained in its database

Compared with conventional small-molecule drugs, products derived from a biological source are structurally complex, large molecules and involved in different manufacturing processes that require tight control to ensure their safety, purity, quality, and efficacy. Because of these differences in respect of the product characteristics and manufacturing process,

The database must also be able to, “accept, store, and manage molecular physiology data” (Moore, 2007). In doing so, selecting the correct database is essential, and many companies choose to create their own private database. However, the creation and use of public databases are on the rise. After all the genetic information has been inputted into a database, it needs to be analyzed. Data analysis consists of data mining and machine learning methods, through the use of high-powered computers. Data analysis is also used for identifying any bimolecular patterns in the observed outcome (Moore, 2007). In addition, data mining is then used to manage and study the genetic information. After all of the analysis and studying, the interpretation of the information is what is most important. The researchers need to be able to understand and comprehend what the data information signifies. From there, bioinformatics can then be used for other various resources. For instance, bioinformatics can be valuable in delivering therapeutics (Bioinformatics, 2000).

Drug research is connected to a range of academic studies such as biology, pharmacology, medicinal chemistry and toxicology. Pharmaceutical researchers can design novel therapeutic drugs based on these studies above. The invention of new drug can be divided by function into two stages: drug discovery and drug development. Drug discovery is the process by which a new drug candidate is found and identified. Distinctively, bringing a new drug candidate to the market through clinical trials is called drug development. The first part of this essay provides an overview of drug discovery and pre-clinical research and development

Introduction AstraZeneca PLC (AstraZeneca, AZN:NYSE, AZN:LSE) is one of the largest pharmaceutical companies in the world. It was formed in 1999 from the merger of Sweden’s Astra AB and UK’s Zeneca Group plc. Core Activities AstraZeneca is engaged in the discovery, development, manufacturing and marketing of prescription pharmaceuticals and biological products for important areas of healthcare: Cardiovascular, Gastrointestinal, Infection, Neuroscience, Oncology, and Respiratory and Inflammation. One of the key benefits of the merger between Astra and Zeneca is seen as their portfolio of new products in development: AstraZeneca call this their 'product pipeline'.

One of the challenges in drug administration and drug development is the need to identify and avoid potential drug–drug interactions (DDIs). During the course of drug development, it is vital to identify possible drug interactions with the new drug compound that is being investigated. Revealing the mechanism behind the drug interactions can provide the necessary scientific evidences for further investigation of the drug compound. The ability to find DDIs and their scientific evidences efficiently and economically can have an impact on the current approach in drug administration and drug development. The reporting systems developed by the Government present an opportunity to monitor drug safety in a large and diverse population of patients. Quantitative signal detection algorithms use these data to flag and prioritize drug-event signals for follow-up analysis via formal pharma co-epidemiological

The three steps route is mostly adopted, in which, the first step the researcher gathers the vital data required for the QSAR process. This can be achieved through collecting, or designing a training set of chemicals that can help obtain, taking relevant and quality data. The second step is to choose descriptors that can accurately relate chemical structure to biological activity since not all molecular descriptors are useful for a particular model. The last step is to apply statistical methods that correlate changes in structure with changes in biological activity (Perkins, 2003). Different modelling techniques are adopted to come up with models that vividly describe the relationship between the structure and the property the researcher is interested in. An example of this are but not limited to, multiple linear regression, logistic regression, and machine learning methods. The validation process ensures that the model obtained is appropriate and useful. In fact, only one-fifth of the molecular descriptors is used to avoid over-fitting the model. In most cases, multiple linear regression is adapted to generate QSAR Models. The model is thus expressed as follows:

Although R&D has been retained by the large pharmaceutical firms, there has been a continuous decline in the R&D productivity. Controlling R&D is imperative to the success of a Pharmaceutical firm. However, as the pharmaceutical industry is maturing, there are diminishing returns to the R&D investment. Fewer and fewer blockbuster drugs are being discovered and therefore R&D is not the most value adding component in the value

Bioinformatics is the applications of computers, software, and mathematical models to the management of biological information from large data sets. It is the storage and analysis of biological information by the use of computers. Computers are used to gather, store, analyze, and integrate information that can then be applied to gene-bases drug discovery and development. This has become an essential part of biological research activity, and is rapidly becoming as ordinary in biosciences as molecular biology has been since the 1980s. The science of Bioinformatics, which is the melding of molecular biology with computer science, is essential to the use of genomic information in understanding human diseases and in the identification of new molecular targets for drug discovery (What is bioinformatics). The primary goal of bioinformatics is to increase the understanding of biological processes. It is the "determination of the sequence of the entire human genome (approximately three billion base pairs)” (What is bioinformatics). This is done so differently from other approaches, however, due to its focus on developing and applying computationally intensive techniques to achieve this goal. "With bioinformatics, scientists can compare data from genetic material of a variety of living things, from tiny bacteria to large organisms, such as humans (Bioinformatics, 2012).