Growing resistance to antibiotics and other antimicrobial treatments is a serious global healthcare challenge. A new study in Antibioticsdemonstrates a method for tracking the spread of genes for antimicrobial resistance among bacterial populations over time. The new computational technique relies on the rapidly increasing availability of bacterial genetic sequences in public databases such as GenBank.

“Our idea is that this could be used as a monitoring system,” explains Ivan Erill, professor of biological sciences at UMBC and the study’s senior author. “It’s great for studies trying to find insight into what’s happening in bacterial genomes.”

The code Erill and colleagues Miquel Sánchez-Osuna and Jordi Barbé at the Universitat Autònoma de Barcelona developed can analyze the sequences of all known bacterial plasmids (little circular pieces of DNA that can exchange genes between bacteria) in about an hour. The results reveal which resistance genes are spreading most and the genes’ likely origin. 

A computational analysis like this is much faster and less expensive than complex systems involving coordination among clinicians around the world. This means it could be carried out more frequently to help doctors and researchers stay updated on shifting resistance threats.

“There’s going to be more and more data that you can mine this way,” Erill says, noting that the amount of genetic sequence data available is doubling approximately every two years. He adds, “I love it because it’s simple. It’s fast, and you can deploy it in a flash.”

Genetic detective work

So how does this new technique work? Microbial DNA, like all DNA, is made up of four bases: A, T, G, and C. A pairs with T, and G pairs with C. However, the ratio of the bases varies considerably across microbial species. Some bacteria are split 50-50 between AT and GC pairs, while other bacterial genomes may contain anywhere from 30 to 70 percent GC pairs. In a previous study, Erill and colleagues leveraged this variability to investigate the emergence of resistance against sulfonamides, an early class of antimicrobials.

As resistance genes hop from species to species via plasmids, they largely retain the GC ratio of their original source. So, if there is a mismatch between the GC ratio of the resistance gene and the rest of a bacterium’s genome, that means the resistance gene has come from elsewhere. The simplicity of this technique means it is not only faster than clinical methods at tracking the movement of resistance genes, but also faster than other computational methods.

If a resistance gene has been in a species long enough, its genetic sequence may eventually begin to approach the GC content of its new host, but that could take millions of years. “For what we’re looking at, which is gene movement in the last 60 to 100 years,” Erill says, “it’s basically a snapshot.”

Specialists spread faster

Using the new monitoring technique, the study authors confirmed that resistance genes are most likely to spread if they are on conjugative plasmids, a type of plasmid that can easily transfer between bacterial cells. Researchers generally already understood this, but confirming it with the new method helped verify the technique’s efficacy.

The new study also found that resistance genes effectively targeting very specific antibiotics spread the most. These genes generally require so many mutations to evolve that they are unlikely to have arisen naturally in any given bacterium since humans started using antibiotics. But if they are present anywhere in the bacterial population when the corresponding antibiotic is introduced, they will spread quickly.

“As soon as there is selective pressure from that antibiotic, there is selective pressure to move this thing around, because it is a bacterium’s silver bullet against that antibiotic,” Erill says. In contrast, generic resistance that requires only a few mutations to existing genes is less likely to spread rapidly, Erill explains. “There isn’t a lot of selective pressure to pass it along, because by the time it comes, the bacterium has likely already discovered it,” he says.

Hospitals aren’t likely the culprit

The new study also found that genes for resistance to antibiotics used in livestock or prescribed outside of hospitals were likely to spread through the global bacterial population. Resistance to antibiotics used in more limited settings hardly spread at all. “That tells you that if you use things cautiously, then there is not so much selective pressure,” Erill says.

Perhaps most important for antibiotic policy moving forward, Erill’s team found that most of the resistance genes came from a single source and then spread, rather than evolving independently multiple times. “Resistance is in the environment,” Erill says, explaining that it needs a vehicle to get into the mainstream.

If antibiotics were only used in hospitals, rather than in livestock and other environments, resistance would be much less common, Erill argues. This is because resistance from hospital use alone “would presume that you have naturally resistant bacteria living in the hospital already, ready to pass on their genes,” he suggests. While there are certainly infectious microbes present in hospitals, “most of the microbial diversity is in the soil and the water,” Erill says. If antibiotics never reach the cells that happen to be resistant, that resistance won’t spread.

While the new study is “a methods paper more than a results paper,” Erill says, “we believe it’s an important contribution.” It puts forward a process for continually monitoring shifts in bacterial genomes over time, which could influence future antibiotic development research or treatment regimens. Perhaps it could even encourage limits on antibiotic use in agriculture and other settings where the drugs can end up in the environment.

Best of all, other research teams can use the new method to pursue answers to their own questions, Erill explains. “You can use it with a very fine comb to poke at whatever you are interested in.”

Written by Richard Forno, principal lecturer in computer science and electrical engineering, UMBC

The Biden administration released its first National Cybersecurity Strategy on March 2, 2023. The last version was issued in 2018 during the Trump administration.

As the National Security Strategy does for national defense, the National Cybersecurity Strategy outlines a president’s priorities regarding cybersecurity issues. The document is not a directive. Rather, it describes in general terms what the administration is most concerned about, who its major adversaries are and how it might achieve its goals through legislation or executive action. These types of strategy statements are often aspirational.

As expected, the 2023 Biden National Cybersecurity Strategy reiterates previous recommendations about how to improve American cybersecurity. It calls for improved sharing of information between the government and private sector about cybersecurity threats, vulnerabilities and risks. It prescribes coordinating cybersecurity incident response across the federal government and enhancing regulations. It describes the need to expand the federal cybersecurity workforce. It emphasizes the importance of protecting the country’s critical infrastructure and federal computer systems. And it identifies China, Russia, Iran and North Korea as America’s main adversaries in cyberspace.

However, as a former cybersecurity industry practitioner and current cybersecurity researcher, I think that the 2023 document incorporates some fresh ideas and perspectives that represent a more holistic approach to cybersecurity. At the same time, though, some of what is proposed may not be as helpful as envisioned.

Some of the key provisions in the current National Cybersecurity Strategy relate to the private sector, both in terms of product liability and cybersecurity insurance. It also aims to reduce the cybersecurity burden on individuals and smaller organizations. However, I believe it doesn’t go far enough in fostering information-sharing or addressing the specific tactics and techniques used by attackers.

The end of vendor indemnification?

For decades, the technology industry has operated under what is known as “shrink-wrap” licensing. This refers to the multiple pages of legal text that customers, both large and small, routinely are forced to accept before installing or using computer products, software and services.

While much has been written about these agreements, such licenses generally have one thing in common: They ultimately protect vendors such as Microsoft or Adobe from legal consequences for any damages or costs arising from a customer’s use of their products, even if the vendor is at fault for producing a flawed or insecure product that affects the end user.

In a groundbreaking move, the new cybersecurity strategy says that while no product is totally secure, the administration will work with Congress and the private sector to prevent companies from being shielded from liability claims over the security of their products. These products underpin most of modern society.

Removing that legal shield is likely to encourage companies to make security a priority in their product development cycles and have a greater stake in the reliability of their products beyond the point of sale.

In another noteworthy shift, the strategy observes that end users bear too great a burden for mitigating cybersecurity risks. It states that a collaborative approach to cybersecurity and resiliency “cannot rely on the constant vigilance of our smallest organizations and individual citizens.” It stresses the importance of manufacturers of critical computer systems, as well as companies that operate them, in taking a greater role in improving the security of their products. It also suggests expanded regulation toward that goal may be forthcoming.

Interestingly, the strategy places great emphasis on the threat from ransomware as the most pressing cybercrime facing the U.S. at all levels of government and business. It now calls ransomware a national security threat and not simply a criminal matter.

Backstopping cyber insurance

The new strategy also directs the federal government to consider taking on some responsibility for so-called cybersecurity insurance.

Here, the administration wants to ensure that insurance companies are adequately funded to respond to claims following a significant or catastrophic cybersecurity incident. Since 2020, the market for cybersecurity-related insurance has grown nearly 75%, and organizations of all sizes consider such policies necessary.

This is understandable given how many companies and government agencies are reliant on the internet and corporate networks to conduct daily operations. By protecting, or “backstopping,” cybersecurity insurers, the administration hopes to prevent a major systemic financial crisis for insurers and victims during a cybersecurity incident.

However, cybersecurity insurance should not be treated as a free pass for complacency. Thankfully, insurers now often require policyholders to prove they are following best cybersecurity practices before approving a policy. This helps protect them from issuing policies that are likely to face claims arising from gross negligence by policyholders.

Looking forward

In addition to dealing with present concerns, the strategy also makes a strong case for ensuring the U.S. is prepared for the future. It speaks about fostering technology research that can improve or introduce cybersecurity in such fields as artificial intelligence, critical infrastructure and industrial control systems.

The strategy specifically warns that the U.S. must be prepared for a “post-quantum future” where emerging technologies could render existing cybersecurity controls vulnerable. This includes current encryption systems that could be broken by future quantum computers.

Where the strategy falls short

While the National Cybersecurity Strategy calls for continuing to expand information-sharing related to cybersecurity, it pledges to review federal classification policy to see where additional classified access to information is necessary.

The federal government already suffers from overclassification, so if anything, I believe less classification of cybersecurity information is needed to facilitate better information-sharing on this issue. It’s important to reduce administrative and operational obstacles to effective and timely interaction, especially where collaborative relationships are needed between industry, academia and federal and state governments. Excessive classification is one such challenge.

Further, the strategy does not address the use of cyber tactics, techniques and procedures in influence or disinformation campaigns and other actions that might target the U.S. This omission is perhaps intentional because, although cybersecurity and influence operations are often intertwined, reference to countering influence operations could lead to partisan conflicts over freedom of speech and political activity. Ideally, the National Cybersecurity Strategy should be apolitical.

That being said, the 2023 National Cybersecurity Strategy is a balanced document. While in many ways it reiterates recommendations made since the first National Cybersecurity Strategy in 2002, it also provides some innovative ideas that could strengthen U.S. cybersecurity in meaningful ways and help modernize America’s technology industry, both now and into the future.

This article is republished from The Conversation under a Creative Commons license. Read the original article and see more than 250 UMBC articles available in The Conversation.

Tolulope Ale, doctoral student, will present his research on studying the anomalous relationships of atmospheric variables of Greenland at the 2023 CRA_WP Grad Cohort Workshop for IDEALS in Honolulu, Hawaii which will take place March 23 - 25, 2023.

Presentation Topic: Multi-domain relationship heterogeneous temporal data

Mentor: Dr. Vandana Janeja --  Professor,  iHARP Director,  and UMBC Department of Information Systems Chair

To find out more about the workshop, please click here.

Tolulope (Tolu) Ale, MSc., is a doctoral student in the information systems department. He obtained a master's degree in physics from the University of Texas at El Paso in 2021 with a focus on computational biophysics. He is interested in applied Data Science and Machine Learning. Tolu is conducting research in the Datamining/Machine Learning area under Dr. Vandana Janeja. His interest is in identifying anomalous associations in temporal data that can be further used in predictive analytics. Tolu's research is focused on mining multivariate anomalous temporal events in the Arctic domain.