Healthcare (Commonwealth Union) – Genomic surveillance plays a crucial role in a wide variety of different fields. These include the monitoring infectious diseases, antimicrobial resistance, tracking biological weapons and many more.
When it comes to tracking and analysing the genetic sequences of pathogens, genomic surveillance is a crucial method for identifying emerging viral threats. However, current global monitoring systems are often expensive, unevenly available across regions, and too slow to detect potentially harmful variants before they spread across borders, increasing the risk of future disease outbreaks.
A new study published in Nature Communications, co-authored by Patricia Ning, an assistant professor of statistics, and Jifan Li, together with researchers from several international institutions, presents a new framework designed to address these challenges. The approach aims to make genomic surveillance faster, more affordable, and more efficient for detecting emerging COVID-19 variants.
Ning’s algorithm is designed to enhance community-level surveillance capabilities worldwide, helping regions improve early detection and response efforts in preparation for future pandemics.
Ning indicated that they wish to make real-time predictions of what will take place so they have the ability to provide guidance for decision-making, for government agencies in particular.
She further pointed out that it can assist in enhancing disease-control efforts.
Ning created a new approach known as the Iterative Block Particle Filter algorithm, which builds on the traditional particle filter technique, also referred to as the Sequential Monte Carlo (SMC) method. The SMC method is a widely adopted computational approach used in many areas of scientific research.
However, the SMC method has a key drawback: it is generally limited to systems with small to moderate spatial or network dimensions. As the complexity and scale of a system increase, the method becomes much harder to apply efficiently.
Although larger networks can sometimes be divided into smaller sections to allow the SMC method to function, genomic surveillance data sets are typically so large and intricate that this strategy is no longer practical or effective.
“For example, we don’t want to disrupt interactions between cities,” Ning said. “When someone flies from New York to California, there is still some interaction. We do not want to break that; we want to preserve it instead of separating cities into isolated study groups.”
The algorithm of Ning was developed to enable scalable inference while overcoming the hurdle of dimensionality — a major challenge in computational methods where errors and processing demands increase exponentially as the number of variables involved grows.
The approach tackles this challenge through an iterative framework, where the results generated from one dataset become the starting point for the next. Instead of relying on a single large-scale global model, Ning’s method performs localized updates while retaining essential connections between nearby regions and major transportation hubs. This design allows the algorithm to scale efficiently by limiting filtering errors within local areas. It is specifically built to handle complex high-dimensional spatial-temporal datasets while preserving relationships between different geographic units.
In the study, the algorithm was tested using large-scale, multi-strain models based on real-world data, including epidemiological information, vaccination records, and detailed global air travel patterns. The models tracked multiple viral strains across dozens of regions. Compared with existing widely used filtering algorithms, Ning’s approach achieved better performance and shortened the time required between identifying a disease variant and completing genomic sequencing — the process of determining the precise genetic makeup of a virus or organism.
The findings suggest that Ning’s algorithm could help governments and healthcare systems improve outbreak preparedness by directing genomic surveillance resources more effectively, particularly toward key international travel hubs. This approach could be especially valuable in regions with limited resources, where maintaining large-scale sequencing operations over extended periods may be challenging.
