The Zika virus circulated in many regions of the Americas for several months before cases of infection were detected, according to new data from an international research team from the Broad Institute of MIT and Harvard and several collaborating institutions.

These findings, revealed in the journal Nature in a paper led by Pardis Sabeti of the Broad Institute and Harvard University, arise from an analysis of 174 Zika virus genomes — including the largest collection of new Zika virus genomes to date — sequenced from patient and mosquito samples collected in 11 affected countries and territories.

The genomic data allowed the research team to reconstruct for the first time the spread of the virus across South and Central America, the Caribbean, and into the southern United States.

In many of these regions, the virus circulated for months before local cases of infection were detected. Sabeti and colleagues’ analysis suggested that Zika was circulating in Brazil around February 2014, a year before that nation’s first confirmed infections were reported. Similarly, the virus appears to have arrived in Colombia, Honduras, Puerto Rico, and elsewhere in the Caribbean from 4.5 to 9 months before the first confirmed local infections, highlighting the importance of having sensitive and specific diagnostic tools early in an outbreak.

These results appear only now, months after the peak of the outbreak, because sequencing Zika virus has proved to be challenging, particularly directly from patient samples. The difficulty arises because Zika virus is typically present at very low levels in patients and disappears quickly. As a result, very few Zika genomes had been generated prior to this study, leaving researchers with little basis for understanding how the virus is spreading and evolving.

To address this lack of data, the team developed new laboratory and analytical methods for capturing robust Zika genomic data, and applied them to samples collected in partnership with collaborators in Brazil, Colombia, the Dominican Republic, Honduras, Jamaica, Puerto Rico, Massachusetts, and Florida to generate 110 new genomes for this study. The team combined those genomes with an additional 64 available in GenBank and in one of the study’s two companion papers to carry out their analysis.

“We knew it was important to understand the viral populations driving the epidemic, which motivated us to tackle the challenges of sequencing Zika,” said study co–first author Hayden Metsky, a graduate student in the Sabeti lab. “Because the data we generated capture the geographic diversity of the virus across the Americas, they provide an opportunity to trace how and when the virus spread. Our data and findings will also support development of more effective molecular diagnostic tests, as well as improved public health surveillance tools.”

The work also highlights the importance of quickly creating trusted partnerships between researchers and across institutions and regions, and of sharing data openly during outbreaks.