Unlocking human genome’s secrets is the greatest data science challenge

27 Oct 2017

Image: Piotr Zajda/Shutterstock

Data science is gathering pace – not just in the technological world, but in the very fabric of our being – resulting in a new age of medicine and drug discovery.

In our effort to find the answers to some of the most afflicting conditions and diseases known to science, biologists are rapidly turning to mapping the human genome to help us solve some of the great questions, such as: why do some people get certain diseases while others don’t?

This story began nearly 30 years ago when biologists and data scientists teamed up to map the human genome for the first time. It would take 13 years of time-consuming and expensive research to finally achieve one of the great milestones of science.

Back then, it cost a total of $2.7bn but now, in almost the same amount of time it took to create one genome, we have more than 1m genomes mapped and it costs as little as $1,000 to generate.

Incredible volumes of data

This, according to Jim Sullivan, global VP of discovery at pharma giant AbbVie, has resulted in modern medicine quickly becoming a convergence of science, technology and big data, pushing researchers to unprecedented areas.

Contemplating the sheer scale of how much data is produced just by mapping a single genome is mind-boggling. For example, if you printed out one human genome on sheets of paper, it would fill a stack of books that would tower 150ft, nearly as tall as the Arc de Triomphe in Paris.

“It’s an enormous amount of data,” Sullivan said, “especially when you think about the software and hardware needed to store all of that data. Fortunately, we’re coming to a point now where we’re figuring that out.”

Jim Sullivan

Jim Sullivan, global VP of discovery at AbbVie. Image: AbbVie

Ireland a great place for genomics research

One such way AbbVie and Sullivan are figuring that out is with the recent collaboration with Genomics Medicine Ireland (GMI), an Irish life sciences company leading large-scale, population-based genome research studies on the island of Ireland, examining the relationship between genetics, health and disease.

Founded in 2015, GMI earlier this year announced it was to partner with AbbVie and global contract genomics organisation WuXi NextCODE on a 15-year project to sequence 45,000 genomes from volunteer participants across Ireland, to seek new insights into the biological processes that underlie complex diseases.

Sullivan, a Meath native who has spent many years living and working in the US, believes his home country is a prime location to develop data-driven genomics research.

“Ireland is a great place to do this work because, in general, there’s a large homogeneous population as well as the fact there’s a clinical network established through major medical centres for patient consent and medical records curation, particularly in Dublin and the major cities,” he said.

Role of AI in the coming years

In the first four to five years of this partnership, Sullivan hopes that quality data science will help it and the other two parties involved to better understand the genetic underpinnings and the genes that are driving many health conditions, and explore ways that could be helpful in clinical trials.

One possible boon to our efforts to map the human genome could come through the use of artificial intelligence (AI).

Just as AI is being recruited to help NASA scientists better understand and find links between vast amounts of data, AI will increasingly be used to help research teams potentially find cures or treatments for diseases such as Alzheimer’s and cancer.

While describing it as an “absolutely powerful technology”, Sullivan said the idea that AI will simply be able to replace and outperform human scientists won’t be the case for the foreseeable future.

“Over time, I think AI is going to help us make better choices in drug discovery into which proteins are most important in disease, which molecules do we think we should make etc,” Sullivan said.

“They won’t replace humans because there’s a huge amount of intuition involved in drug discovery, but it will have a very important enabling function as we do our drug discovery and development.”

A lot changes in 25 years

Having led more than 100 compounds into clinical development across a number of disease states and technology platforms for nearly 30 years, Sullivan is amazed at the pace of change in his field.

One notable difference is that the make-up of a typical pharmaceutical lab is now much more diverse in terms of the roles that are involved in finding the next generation of drugs.

“The typical grouping of scientists 25 years ago in drug discovery were chemists and biologists,” Sullivan said.

“Now, we have data scientists, bioinformationists and computational biologists, all of whom are helping us look at the enormous amounts of data we’re generating from these human genetics collaborations as well as our own internal work.”

Where data science will take genetics and drug discovery in the next 10, 20 or even 50 years seems about as possible as trying to predict the weather – and that’s a good thing.

Jim Sullivan is the global VP of discovery at AbbVie and will be attending BioPharma Ambition, to be held in Dublin on 21 and 22 February 2018. BioPharma Ambition is a global conference showcasing innovative medical, research and training advancements in healthcare.

Colm Gorey was a senior journalist with Silicon Republic

editorial@siliconrepublic.com