Wellcome Trust big data analytics symposium

Cambridge, UK. - Across the life sciences, statistics and analytics are the crucial elements in understanding and capitalizing on the genomics revolution and in driving forwards personalized medicine. That is the message left with the attendees of the Wellcome Trust Genome Campus symposium on Big Data Analytics for Genetics in Personalized Medicine, on Saturday 9th September.

Clinicians, doctors, researchers and students gathered at the Francis Crick Auditorium in the bright, early autumn sun to await the day’s talks by some prominent experts from world renowned institutions.

The presentation that created the most buzz on the day was a rallying call for the need to crowd-source and globally collaborate on variant curation from Broad Institute Member, Dr. Heidi Rehm, who spoke about standards and approaches to support genomic variant interpretation. This has recently become tragically topical as, in the US, there has been the first lawsuit over a misdiagnosed variant.

Dr. Rehm told the audience: “We are seeking to redefine clinical relevance and disease validity clarification through a semi-quantitative system, aggregating clinical evidence and giving variants a validity score. At the moment, too many genes which are reported as related don’t have enough supporting information. What we do have is a huge number of experts from whom we can essentially crowdsource this. The database is open to the public and follows all ACMG rules and uses industry accepted ideas and terminology. Submissions have to show how they came to their conclusions and then we work as a community to resolve conflicts and discordance. In fact, so far, we’ve resolved 72% of the 300 conflicting variants we’ve faced so far. Systematically we can reassess outliers, support clinicians and affect patients.”

The next came from University of Southern California’s Associate Professor in the Department for Preventative Medicine, Dr. Kimberly Siegmund who spoke about using statistical approaches to investigate changes in the mutational processes during the growth and development of a cancer or tumour.

She said: “We know tumours share ancestry, that they all have a tumour initiating event. We can learn about the mutation process and sequence tumours to build a catalogue from which we can then infer patterns and draw contextual data from neighbouring bases. This gives us multinominal probability.”

Dr Seigmund’s lab undertakes multi region sampling and whole exome sequencing with a minimum of three runs to understand the diseases trunk and branch mutations and try to glean any de novo mutations contained within to add to the catalogue using a descriptive framework.

Underpinning these ideas was a phrase from the day’s first speaker, Dr. Chris Spencer, head of innovation at Genomics PLC who spoke generally on improving healthcare with genomics. He said: “Genetics is the best chance in a generation to make a major shift to better healthcare, at the very least, to improve the efficiency of the doctor’s toolbox. Over the next 10 years, as a global industry, we will be sequencing millions if not billions of genomes but unless we make use of that data it will be useless. Frankly, even if all the sequencing machines in the world blew up this afternoon, we would still have enough data to be working on for the next 50 years!”

After the intermission, things took a turn to the technical. First, attendees heard from Dr. George Davy Smith, Professor of Clinical Epidemiology at the University of Bristol, UK, who spoke about Mendelian randomisation in a data rich environment.

Dr. Davy Smith utilized a number of formulae and graphs focusing on utilising resources such as the UK Biobank to more deeply understand factors of disease progression and how pleiotropy can influence outcomes but that variants may still be data derived.

The last speaker of the day was Dr. Hae Kyung Im, Assistant Professor of Genetic Medicine at the University of Chicago, USA, who spoke about integrating GWAS, omics and electronic medical records to dissect disease biology.

“Recent progress has allowed us to generate huge amounts of HTS data and to be able to interpret this data. At Chicago we are ambitious and we want to characterise the phenotypic consequence of all human genes.” She said.

Dr. Im went on to talk about her new project, The Human Knockout Project, which is a systematic effort to understand the disruption of any given genes and to learn more about why some drugs do not work on some groups.

The day overall was a great success. Complemented by the lovely weather, the conducive staff and inspiring exhibitions attendees and speakers alike enjoyed a stimulating and interesting range of discussions and debates.