Pardis Sabeti is a computational geneticist who works at the Broad Institute of MIT Harvard, where she pinpoints the genetic factors in genomes that allow diseases to take hold.
When the Ebola outbreak struck in 2014, Sabeti worked between a clinic in Kenema, Sierra Leone, and her lab in Harvard to sequence the virus's genomes. Her lab mined the samples for clues about how it evolves and spreads. As they worked, they put the data online, free to use, to help drive the development of diagnostics and treatments. At the time, the first batch of 99 genomes sequenced made up the biggest published dataset on Ebola.
Sabeti has become an advocate for sharing data online to speed up response. Here she tells WIRED about how Ebola survivors have become a focus of her research, and a new open-access app that will improve diagnosis of infected patients.
WIRED: What are your lab's central aims?
Pardis Sabeti: One of our strengths is computational biology: taking genomes and trying to understand what's meaningful and what changes have been important. I look at the human and viral sides, and both have a lot to offer - for instance, tracing the human genome to understand what were the major pressures on it, and in doing so, pointing to things we didn't know about that are important. We're mining the genome for clues.
Why gather data on viruses such as Ebola or Lassa?
Everything we do is based on the genome sequences of a virus. All sorts of data need to be available - clinical data, genomic data, epidemiological data. But when it comes to genome data, you need it for everything: the diagnostics; the vaccines; and the therapies. These are our major action items in the lab.
A major focus for you has been making genetic data on Ebola freely available. Has this led to any breakthroughs?
Of the things I've done in my career that have made me most proud, it was when we released this data. Many people have since told me how they have developed a new diagnostic based on our sequences1. Knowing that was made possible was a big deal. Some vaccines were developed using it too. Vaccines and diagnostics are practical things which we know more data will allow to be as good as they can be.
What's the status of your genome-sequencing work?
It's continuing. As new cases emerge we're sequencing them to understand where they might have come from and how they might relate to others. Also, there is an interesting case of individuals who are likely reservoirs for the virus. They have Ebola in their semen and in their breast milk, and that's frightening. People will say: "Oh, they survived Ebola," but with both Lassa and Ebola the long-term effects on health are dramatic. It's something we need to work on to stop. We're trying to understand what the viral population is in these people and how it is changing.
You've also built an app called Ebola CARE that predicts patients' outcomes based on symptoms. How does it work?
Using a small clinical data set from Ebola patients during the West African outbreak, we developed a statistical method that's expected to help clinicians predict a new patient's prognosis - their probability of fatality or recovery - to nearly 100 per cent accuracy2. To make it available in the field we've made the mobile app and the algorithms behind it open and flexible to being updated with better data, to encourage others to make outbreak data open access. If we can create tools that give clinicians real-time information, they're likely to put data into a system that other people can access. They can help drive research.
You've also been working on the Zika virus. Is this building on your Ebola research?
The technologies we've been developing for Ebola work well for many viruses, but Zika is more challenging as it often occurs at very low concentrations in patient samples. Read more: Zika: how can it be stopped?
For this reason, researchers have only been able to sequence a small number of samples over the past year, leaving many questions about the virus's evolution and diversity.
My lab has been developing technologies to sequence the Zika virus.
The technology is ready; our challenge is finding collaborators. We welcome anyone who'd like to work with us.
Do you think the push for transparency and data sharing during the Ebola epidemic has changed the conversation around outbreak response?
There's definitely been a sea change, with a great deal of messaging now about open access and sharing data. But we need to work to make sure that we follow this not just in our language but in our actions and beliefs. These situations in an outbreak can be fast-moving and we are in a race against time. We need to figure out how to move things forward, because this was a major warning to the world - but it's not the worst we have or could see.
\1. Stephen K. Gire, 2014. "Genomic surveillance elucidates Ebola virus origin and transmission during the 2014 outbreak," Science 6202, 1369-1372
\2. Andres Colubri, 2016. "Transforming Clinical Data into Actionable Prognosis Models: Machine-Learning Framework and Field-Deployable App to Predict Outcome of Ebola Patients," Public Library of Science Neglected Tropical Diseases
This article was originally published by WIRED UK
