GlaxoSmithKline Takes Drug Discovery To The Next Level

Dr. GP Pulipaka
Jul 22, 2017 · 7 min read

Big data could be the future of the new antibiotic in the health care industry to enhance the patient medical care. Big data is shifting the paradigm by creating a new data-driven health care industry perfecting the precision on performing the surgery based on the data points built by big data analytics. Big data analytics tools are aiding corporations in developing new medicine every day with data science more accurately. According to a modern medical research estimate, the pharmaceutical product cost from the lab to the consumer is around $2.6 billion dollars. McKinsey, research estimated big data analytics tools would lower the health care industry costs to the tune of $493 billion dollars. The challenges to invent a new drug arise from experimental clinical trial costs and data integration between pharmaceutical companies for performing clinical trials. Several data sets create a common functionality between multiple pharmaceutical companies. However, the medical data is generated from heterogeneous sources of structured and unstructured big data such as paper-based health records and electronic medical records. The divergent database systems do not have similar formats to share the data between health care networks. Therefore, usage of electronic medical records aids the health insurance providers, clinics, and health care consumers to port the data quickly. According to a research study conducted by Center for Disease Control (CDC), there is a surge of 78% in the usage of universal electronic medical records in 2013 contrasted with 18% usage in 2001 by the hospitals and clinics in bricks and mortar environment.

Big data can be a boon or a curse to a clinic, hospital, health care provider, and health care insurer. SAS built big data analytics tool for GlaxoSmithKline (GSK) running on a convergent cloud connecting Big Pharma companies in the medical industry to divvy up and distribute the clinical trials big data among companies. This collaboration is a new revolutionary advancement in the health care industry to scale down the expenditure mixed with a myriad of factors on research and development of a new medicine in the medical field. GSK made data sets available from 200 anonymous clinical trials to every Big Pharma company to utilize the data for medical advancements. However, there have been ethical concerns about sharing the electronic health records and clinical trials data from Health Insurance Portability and Accountability Act (HIPPA). HIPPA laws make it abundantly clear that personal health information must be maintained with high security and privacy for protecting the patient privacy. The privacy laws apply to the medical data stored on the mobile devices as well. In a recent data breach, health insurance provider Anthem was hacked. 80 million personal data records released into the market. However, Anthem clarified, protected health information was in compliance with HIPPA. While big data made strides in revolutionizing health care industry, 40 million Americans were affected by the data breach of protected health information through various health insurance providers, clinics, and hospitals through the end of 2014. The big data from health care industry is generating unprecedented amounts of data. The health care industry has seen an alarming rate of 25% surge in the data breaches related to electronic medical records and HIPPA’s classified protected health information of the patients.

Figure 1. GSK, NCI, and DOE ATOM Consortium for HPC drug discovery through deep learning. ATOM Consortium

Protected health information data resides with health care providers, health care insurers, clinics, and hospitals throughout the lifetime of the patient. The patient electronic medical records can be transferred from one health care provider to another electronically as well as through smartphones and tablets. However, HIPPA requires protection to the data on these devices as well. According to the research conducted by Ponemon Institute, only 54% of the clinics have competencies to shield the protected health information big data. HIPPA laid out privacy laws to prevent the misuse of protected health information of the health care consumers based on Fair Information Practice Principles for purging the paper-based and the electronic medical records, storage media, and electronic media once the lifecycle of medical data defined by HIPPA expires. HIPPA performs an audit that encompasses 77 rules. Pharmaceutical companies, health care insurers, health care providers, clinics, and hospitals have to follow these principles to safeguard the data from clinical trials, electronic medical records of the medical consumers. The physical facilities, software and hardware systems, data, network infrastructure, cloud computing, security, and access to the clinics, hospitals, health care providers, and insurers must be in compliance with HIPPA regulations.

Figure 2. GSK, NCI, and DOE ATOM Consortium clinical trial and drug research development. NCI.

The healthcare industry spends $ 150 billion every year on research and development costs to introduce new drugs into the market. According to McKinsey & Co. Inc., a global management-consulting firm, big data tools can reduce the research and clinical development by $ 40 billion to $ 70 billion. Data is not useful without information. SAS Institute from North Carolina launched an initiative with SAS In-database and SAS High-performance analytics software. It has built a private cloud to store such big data from clinical trials for co-innovation between all big pharma companies who can access the data from the private cloud and reduce the research time for product development. Data is not useful without information. GlaxoSmithKline’s private cloud has moved one of the largest needles in big data movement for the healthcare industry. This program opened the gateway for global pharmaceutical companies for co-innovation.

In 2017, GlaxoSmithKline, National Cancer Insitute, and Department of Energy formed ATOM Consortium (An Accelerating Therapies for Opportunities in Medicine (ATOM). ATOM’s ambitious project is to apply deep learning for drug discovery through high-performance computing and DOE Labs at Lawrence Livermore National Laboratories by scanning through the big data of millions of molecules and identifying the relationships in genomics and apply to the new data. Drug discovery with deep learning frameworks requires profound amounts of large-scale big data to research and apply on molecular structures. Oak Ridge Laboratories that specializes in DANNA (Dynamic Adaptive Neural Network Arrays) neuromorphic architecture inspired by brain-like biological architecture that can compute millions of data points with low-power cost using spiking neural networks collaborates with ATOM Consortium for drug discovery. DANNA dominates many neuromorphic hardware implementations that have fixed number of neurons and static synapses. DANNA networks are programmable with the design of evolutionary optimizations.

Figure 3. DANNA Network architecture. ORNL
Figure 4. Resulting network architecture. ORNL.

Stanford researchers applied machine learning and deep learning techniques with deep neural networks through multi-layer architectures. Stanford researchers have applied One-Shot learning paradigm with long short-term memory on TensorFlow with DeepChem library. GlaxoSmithKline has agreed to share clinical trial data with chemical compounds and toxicology big data for a period of 15 years. GlaxoSmithKline is expected to revolutionize the drug discovery in the next few years with supercomputing and deep learning.

Figure 5. Stanford researchers One-shot learning for drug discovery. ACS Central Science
Figure 6. Stanford researchers One-shot learning for drug discovery. ACS Central Science
Figure 7. Stanford researchers One-shot learning for drug discovery. ACS Central Science
Figure 8. Stanford researchers One-shot learning for drug discovery. ACS Central Science

References

Altae-Tran, H., Ramsundar, B., Pappu, A. S., & Pande, V. (2017, April 3). Low Data Drug Discovery with One-Shot Learning. ACS Central Science, 3(4), 283–293. http://dx.doi.org/10.1021/acscentsci.6b00367

Ausick, P. (2015). Data Breaches Top 600 to Date in 2015. Retrieved October 17, 2015, from http://247wallst.com/technology-3/2015/10/16/data-breaches-top-600-to-date-in-2015/

Berger, D. (2015). Redspin Issues Annual Healthcare Data Breach Report. Retrieved October 17, 2015, from https://www.redspin.com/company/news-and-events/redspin-issues-annual-healthcare-data-breach-report.php

Donoghue, K. (2014). Top 5 US Healthcare Security Risks. Retrieved October 17, 2015, from http://www.peak10.com/top-5-us-healthcare-security-risks/

Freiherr, G. (2015). How Big Data can help save $400 billion in healthcare costs. Retrieved October 17, 2015, from http://www.cio.com/article/2993986/big-data/how-big-data-can-help-save-400-billion-in-healthcare-costs.html

Greenspan, E. J. (2017). Accelerating Therapeutics for Opportunities in Medicine (ATOM) Collaboration. Retrieved July 22, 2017, from https://deainfo.nci.nih.gov/advisory/fac/0517/Greenspan.pdf

HHS (n.d.). . Retrieved October 17, 2015, from http://www.hhs.gov/ocr/privacy/hipaa/enforcement/examples/disposalfaqs.pdf

Hsiao, C. (2014). Use and Characteristics of Electronic Health Record Systems Among Office-based Physician Practices: United States, 2001–2013. Retrieved October 17, 2015, from http://www.cdc.gov/nchs/data/databriefs/db143.htm#x2013;2013</a>

Mullin, R. (2017). Pharma partnership applies deep learning to very big data. Retrieved July 22, 2017, from http://cen.acs.org/articles/95/i4/Pharma-partnership-applies-deep-learning.html?h=-1917987131

Pulipaka, G. (2015). Big Data Appliances for In-Memory Computing: A Real-World Research Guide for Corporations to Tame and Wrangle Their Data (2 ed.). San Bernardino, CA: High Performance Institute of Technology.

Schuman, C. D., Disney, A., & Reynolds, J. (n.d.). Dynamic Adaptive Neural Network Arrays:A Neuromorphic Architecture. Retrieved July 22, 2017, from https://web.eecs.utk.edu/~plank/plank/papers/2015-Schuman-SC-Workshop.pdf

TheWeek (2015). A brief guide to Big Pharma. Retrieved October 17, 2015, from http://theweek.com/articles/583337/brief-guide-big-pharma

Tsai, R. (2014). Sharing of Clinical Trials Data is Moving Forward: What About BioSamples? Retrieved October 17, 2015, from http://blog.fisherbioservices.com/sharing-clinical-trials-data-is-moving-forward-what-about-biosamples

Dr. GP Pulipaka

Written by

Ganapathi Pulipaka | Founder and CEO @deepsingularity | Bestselling Author | Big data | IoT | Startups | SAP | MachineLearning | DeepLearning | DataScience

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade