How AI and Genomics Can Help Fight Antibiotic Resistance

Applying innovative technologies to solve a growing problem.

Posted Nov 15, 2018

geralt/pixabay
Source: geralt/pixabay

Antimicrobial Resistance (AMR) is a global health threat. AMR accounts for more than 700,000 deaths globally each year, and is projected to increase to more than 10 million deaths by 2050, according to a report issued by Wellcome Trust and the UK government. Can applied innovative technologies such as artificial intelligence (AI) and genomics help solve this problem?

The first modern commercialized antibiotic, penicillin, was developed in 1928 by Sir Alexander Fleming. By 1940, the first antibiotic resistance was identified in penicillin-R Staphylococcus. Over time, new antibiotics drugs have come to market, and many of the targeted germs have evolved and developed resistance. AMR is caused by the overuse of antibiotics by humans and livestock, as well as the increased use of antibacterial cleaning and hygiene products. Antibiotics work on bacteria, not viruses, and are often overprescribed for non-viral illnesses. To compound the problem, fewer antibiotics are being developed by pharmaceutical companies due to economic and regulatory barriers, according to the U.S. Centers for Disease Control and Prevention (CDC).

Billionaire philanthropist and Microsoft Co-founder Bill Gates has repeatedly warned that a global pandemic is an existential threat that needs to be addressed.

“If anything kills over 10 million people in the next few decades, it's most likely to be a highly infectious virus rather than a war. Not missiles, but microbes. Now, part of the reason for this is that we've invested a huge amount in nuclear deterrents. But we've actually invested very little in a system to stop an epidemic. We're not ready for the next epidemic.” Bill Gates

Gates uses the example of a pandemic virus as a potential microbe that can wreak havoc. Yet viruses are not the only threat. The lack of antibiotics to treat bacteria-based illnesses present a major global health issue as well. To address this issue, pioneering researchers are using technology to find new solutions.

Recently, research scientists at the University of California San Diego have created a method to identify and predict which genes cause infectious bacteria to become resistant to antibiotics using AI machine learning. The team published its research findings in Nature Communications.

The scientists developed a machine learning computational platform “complemented with genetic interaction analysis and 3D structural mutation-mapping” that can identify “signatures of AMR evolution to 13 antibiotics.” They trained the machine learning algorithm on both genome sequences and phenotypes of 1,595 strains of on tuberculosis-causing bacteria called Mycobacterium tuberculosis.  As a result, the algorithm correctly predicted 33 known antibiotic resistant genes, and identified 24 new genetic signatures of antimicrobial resistance. According to the UC San Diego researchers, their approach can be applied to other infection-causing pathogens.

Researchers at Virginia Tech have developed an AI deep learning solution, DeepARG, to fight antibiotic resistance. Using Next Generation Sequencing (NGS) technologies such as Illumina, DeepARG consists of two models: short sequence reads (DeepARG-SS) and long gene-like sequences (DeepARG-LS). According to the Virginia Tech scientists, “high throughput DNA sequencing technology now provides a powerful tool to profile the full complement of DNA, including ARGs” (antibiotic resistances genes). The researchers have curated the ARGs predicted with “a high degree of confidence” in a database called DeepARG-DB that is available for query or download to support development of antibiotic resistance-related resources.

Day Zero Diagnostics, a venture-capital and angel investment funded startup founded in 2016, is applying whole genome sequencing with AI machine learning algorithm called Keynome™ to shorten the identification of a bacterial infection from two to five days to hours. The company developed a proprietary microbial resistance database called MicrohmDB® that determines antibiotic resistance using genomic data. Day Zero Diagnostics is based at the Harvard Life Lab and works in collaboration with Dr. Doug Kwon at the Ragon Institute of MGH, MIT and Harvard.

According to the CDC, more than two million Americans are impacted with antibiotic-resistant infections and 23,000 die each year from antibiotic-resistant infections. The CDC estimates AMR’s economic impact in the U.S. to be more than $20 billion in direct healthcare costs, plus $35 billion in lost productivity, for a total of $55 billion annually. Through the work of pioneering scientists and researchers, innovative technologies such as AI machine learning and genomics are being applied in hopes of helping humanity in the future.

Copyright © 2018 Cami Rosso All rights reserved.

References

Prestinaci F., Pezzotti P, Pantosti A. Antimicrobial resistance: a global multifaceted phenomenon. Pathog Glob Health. 2015;109(7):309-18.

Wellcome Trust, UK Government. “Review on Antimicrobial Resistance. Antimicrobial Resistance: Tackling a Crisis for the Health and Wealth of Nations. 2014. https://amr-review.org/

Bill Gates. “The next outbreak? We’re not ready.” TED2015. Retrieved 11-15-2018 from https://www.ted.com/talks/bill_gates_the_next_disaster_we_re_not_ready?language=en

University of California – San Diego. “Machine learning identifies antibiotic resistance genes in tuberculosis-causing bacteria.” Phys.org. October 25, 2018.

Kavvas, E., Catouiu, E., Mih, N., Yurkovich, J., Seif, Y., Dillon, N., Heckmann, D., Anand, A.,Yang, L., Nizet, V., Monk, J., Palsson, B. “Machine learning and structural analysis of Mycobacterium tuberculosis pan-genome identifies genetic signatures of antibiotic resistance.” Nature Communications. 17 October 2018.

Gustavo A. Arango-Argoty, Emily Garner, Amy Pruden, Lenwood S. Heath, Peter Vikesland, Liqing Zhang. DeepARG: A deep learning approach for predicting antibiotic resistance genes from metagenomic data. (2018 microbiome).

Day Zero Diagnostics. Retrieved 11-15-2018 from https://www.dayzerodiagnostics.com/

Centers for Disease Control and Prevention. “Antibiotic / Antimicrobial Resistance (AR / AMR). Retrieved 11-15-2018 from https://www.cdc.gov/drugresistance/about.html