Understanding how infectious diseases spread depends on unlocking secrets held in existing data

This applies to infectious disease transmission, main topic of a recent workshop organised by the European Science Foundation (ESF), which called for development of new mathematical and statistical tools capable of probing deeper into existing databases relating to human contact and pathogens.

“One of the most exciting conclusions we came to was the realization that vast amounts of information were already available in various data banks,” said Mirjam Kretzschmar, convenor of the ESF workshop, from the Medical Centre at Utrecht University in the Netherlands.

These databanks held wide ranging information relating to the contact networks of people with various diseases, identifying the pattern of transmission, and also the genotype (total genetic sequence) of the associated pathogens, which could be bacteria or viruses, and potentially in future protozoa or micro fungi. The ESF workshop highlighted the great challenge involved in bringing all this data together, and drawing the correct conclusions from it. “It is clear now that methods are still lacking to unlock the knowledge that is immersed in these data,” said Kretzschmar. “In particular we need to bridge different disciplines to meet this challenge of handling increasingly large amounts of systematically collected genetic sequence data.”

The point here is that particular pathogens such as the influenza virus as identified by their genotype can have different transmission patterns – some might spread faster, and some might require closer contact between people to spread. Therefore correlating details of the unfolding contacts networks between sufferers, as a disease spreads, with the possibly evolving genotype of the pathogen involved can yield valuable insights into the molecular factors relating to transmission. Similarly the disease may become more or less severe over time, or it may affect some types of people worse than others. These factors can also be analysed, helping perhaps to identify what makes a pathogen dangerous and then hopefully predict in advance how an infection may develop and spread. This can lead to appropriate strategies to combat a disease, such as development of a vaccine or public health recommendations, for example that certain high individuals stay at home where possible.

“Research projects have shown that genotyping of pathogens can lead to insight into how risk networks are connected with each other and whether an outbreak consists of many small local outbreaks or whether we are dealing with a supra-national outbreak that requires different intervention strategies,” said Kretzschmar. For example the outbreak of LGV (Lymphogranuloma venereum) among gay men in Europe during 2005 occurred in clusters and analysis showed that clusters from geographically distant cities were connected with each other. Also, the analysis of genotypes showed that the outbreak was connected to earlier cases observed in San Francisco and had been going on for a long time already. This helped coordinate intervention activities among European countries.

The conference also heard how genotyping was helping identify emerging pathogens more quickly than previously, and their likely patterns of transmission. For example genotyping enabled the insect-born virus Chikungunya and its transmission routes to be rapidly identified during a recent outbreak in Italy.

Such information can be correlated with contact tracing networks, which identify the transmission chains taken by an infection. “Two types of contact tracing were discussed,” said Kretzschmar. “One was the traditional way of asking infected persons who their contacts and possible sources of infection were, and the other was tracing clusters of persons who have similar characteristics.”

The latter type of contact network could be identified by taking people with similar genotypes, and also by considering people with similar types of risk behaviour that puts them at risk – gay men already with specific types of high risk behaviour being an example for LGV. “It shows that molecular typing can help uncover risk networks, especially if there is epidemiological information available that allows us to localize these networks and identify persons who are at risk but not yet infected.”

The ESF workshop proved invaluable in identifying the types of data that could increase understanding of how infectious disease spreads and develops in populations, and the statistical and mathematical tools that could perform the required analysis.