NEW YORK (GenomeWeb) – At the Clinical Virology Symposium in Daytona Beach, Florida last week, clinicians and investigators gathered to discuss the latest research on a variety of viruses, as well as their effects on patients. Several presenters spoke about the benefits of next-generation sequencing methods in helping them learn which pathogens they were dealing with in a given patient, or how those pathogens had evolved.
In a presentation of oral abstracts, two researchers in particular — Sergei Belanov of the University of Helsinki and the University of Toronto's Ramzi Fattouh — spoke of their experiences using whole-genome and molecular sequencing.
Belanov presented research published by his group and their collaborators in Genome Biology and Evolution in November, detailing how whole-genome sequencing of the influenza A virus can be used to influence the selection of influenza strains for vaccines. The flu causes epidemics and pandemics every year, but it can be difficult to detect the exact strain responsible, Belanov said. Recommendations from the World Health Organization on vaccine formulation are based on phylogenetic analysis from the previous year, but WGS can be used to analyze the evolution of flu viruses using thousands of samples representing different geographic regions, he added.
The researchers sequenced 3,969 influenza A H1N1 strains and 4,774 H3N2 strains found in public databases from the 2009 through 2015 flu seasons. As a result of the analysis, Belanov said, the team found 481 changes in amino acid substitutions in the H1N1 strain, 61 of which were evolutionary markers that seemed to contain the most information about how the virus evolved. About 40 of these markers circulated during the 2013-2014 flu season, Belanov said, but only two of them were present in vaccine strains for this same season, even though there was a strain from Hawaii which contained all 40 markers.
Similarly, the H3N2 analysis found 533 changes in amino acid substitutions, 68 of which were evolutionary markers. About 33 of these circulated in the 2013-2014 flu season, though only 23 were present in vaccine strains. Belanov and his colleagues, however, found all 33 in a flu strain from Stockholm.
This study, he added, shows that NGS techniques can be used to study flu strains to determine which ones will be most effective as vaccines against the largest number of strains in a given flu season.
Toronto's Fattough, meantime, spoke about the utility of sequencing in outbreaks. He presented research his group has recently done on a suspected outbreak of human adenovirus in hematopoietic stem cell transplant patients in 2015 in Toronto.
Clinicians had observed a cluster of adenovirus among bone marrow transplant patients. After genotyping the patients and comparing them to genotypes of frozen adenovirus samples, Fattouh and his team found eight cases of A31 subtype adenovirus and four cases of C1 subtype. But their findings were difficult to interpret as there were no significant epidemiological links between the patients or even between the patients with each subtype of the virus. So they didn't know where the outbreak had started, or why the patients were exhibiting these two particular subtypes.
They started by considering whether the A31 and C1 types were simply the strains that happened to be circulating in 2015. Looking at studies going back to the 1960s, they found that C1 is among most the prevalent genotypes, but when they genotyped around 30 adenovirus positive specimens from January to August 2015, the researchers didn't find any A31 or C1, so they concluded that the outbreak was not just a matter of prevalent genotypes of virus circulating among patients.
They proceeded to ask for help from their core genomics facility to sequence the hexon and E3 genes of the A31 cases. But as important, Fattouh said, they needed a way to interpret the results correctly to determine what the differences between the virus's sequences meant, and what significance each had. So they built an interpretation framework to interpret the results based on previously reported sequences.
For example, when they looked at a 2006 study of the hexon gene, the researchers found that two unrelated A31 strains could differ by a little as three base pairs. And a 2015 study showed two unrelated A31 could have no differences in hexon and as few as five base pairs of difference in E3.
They sequenced the strains from their outbreak and compared them to three reference strains of A31. They found that hexon and E3 in their samples were somewhat different from those in the reference samples, and are now planning to conduct whole-genome studies to look at relatedness of various A31 strains. They also plan to start looking at C1 strains in the same way, according to Fattouh.
This work demonstrates the value of molecular techniques in defining and mapping the anatomy of an outbreak, he said, as relying on genotyping alone would have sent the team on a wild goose chase as to how the strains of the virus are related. It also demonstrated the paucity of standardized interpretation criteria for relatedness using molecular sequencing for viruses, he added, emphasizing other speakers' calls for increased use of NGS approaches and technology in virology research.