This week we heard from Illumina and European Bioinformatics Institute how consumer access to their own genome for around $500 might change our approach to healthcare.
Illumina is well placed to comment, with ~60% world market share for rapid genome sequencing. The price of a personal genome has dropped from $100M in 2007 to $10,000 today, and their roadmap suggests a price of $1000 within a couple of years, well within the reach of many consumers. Parallelization of sequencing has driven this huge reduction in time and cost, sequencing 1BN fragments at a time, and this allows 40 fold sampling to give 99.999% certainty. Will doctors become mere bystanders as consumers get entirely automated analysis of their inherited diseases? Mark Henderson of the Times was identified as popular journalist doing a particularly good job of laying out some of the social implications of the technology.Illumina has already been trying to make diagnostic use of this data for senior biotech executives to understand the hurdles. Unfortunately, the reference datasets of likely inherited disorders created before the genome was mapped are inconsistent with the current reference dataset. Of over 96,000 genes associated with inherited disorders in Human Gene Mutation Database, only half map onto genes in the reference dataset. For a specific 50 year old biotech CEO Illumina found 16 matches, most of which would indicate death in early infancy, but re-checking against other published papers confirmed all 16 were actually compatible with life. So we are some way from having an entirely automated analysis, although this might quickly change if consumers start donating their personal genomes and their personal health histories to science at death.
Instead, Illumina see the personal genome as just one element helping doctors to treat patients. Cancer mutates away from from our inherited germ line - so identifying differences for specific cancer can reduce time wasted on expensive treatments. An 80 year old with pulmonary metastases was not helped by two treatments of an expensive drug costing $26,000, but by having identified that it was changes in a specific target which characterised this cancer, another drug was effective.
If many consumers start to capture their own genome data, we will need literally astronomical data storage, similar to the volumes of data currently used to analyse star patterns. However, as the EBI speaker pointed out, within our audience carrying iPods and smartphones there was probably around 5 terabytes of data storage available in the room, so it is just a question of gaining access. EBI have gained experience in dealing with large scale genomic data through participation in the 1000 Genomes project, which (by data size) is probably the largest distributed project ever attempted in biology. But few consumers currently pay for their own CT scan (£150) or MRI scan (£300). Institutions trying to contain the cost of malpractice insurance may just want to have tested for some very specific things and then discard the full genome, to protect themselves against being required to produce the data later and getting sued for not having caught other problems. If this happens, it will save a lot of storage, but we will learn a lot less. Tackling the whole genome rather than just trying to characterise known areas found and anotated 30% more genes in the parts that would have been discarded as junk.
Cambridge has a lot of biological, informatics and electronics knowledge to bring to this exciting field - and we also had visitors from Finland, Italy and the USA. Our audience included Amsurg, base4, Cambridge Angels, Eagle Genomics, Judge Business School, Laurea, Microsoft Research, Mills & Reeve, P&G, PA, Pfizer, PHG Foundation, Philips, Plarion, Population Genetics Technology, Somaxa, Swiss Precision Diagnostics, Team Consulting, Wellcome Trust Sanger, TTP, Unilever, and University of Cambridge. Both the technical and the ethical challenges of bringing consumers the benefits of this disruptive technology look set to be keeping them all busy for some time!
Comments