According to the UK's Alzheimer's Society over half of the world's 18m dementia sufferers have Alzheimer's disease.
With the world's ageing population, this number could grow to 34m by 2050.
Funding for Alzheimer's disease research lags way behind some other diseases, with the UK investing as little as £11 pounds for every Alzheimer's sufferer in the country, this compares to £289 of research for every cancer patient.
The new research will scan DNA samples taken from 6,000 sufferers of late-onset Alzheimer's disease and 8,000 healthy control samples from the UK and the US to identify common genetic variations that increase the risk of the disease.
The team will be led by Professor Julie Williams at the University of Cardiff, who will be collaborating with researchers at the Wellcome Trust's Sanger Institute.
"Alzheimer's is a genetically complicated story involving many genes, so we need large sample sizes to make sure any genetic links that we find are not mere coincidence," said Professor Williams.
"With access to 14,000 DNA samples, our study is the largest genetic study ever to look at Alzheimer's and will undoubtedly produce some valuable insights into what causes this devastating illness."
The researchers will search the samples for variations in 500,000 SNP (single nucleotide polymorphisms) genetic markers that can be linked to Alzheimer's disease.
"It's very likely that we will find some unexpected associations.
We know already that certain genes are involved in more than one form of dementia and that even genes that affect cholesterol level can be a risk factor for Alzheimer's," said Prof. Williams.
The genotyping experiments will be conducted by Dr Panos Deloukas' group at the Sanger Institute.
"Alzheimer's disease is a major burden on our society and this burden will only increase as our population ages," says Professor Richard Morris, Head of Neurosciences and Mental Health at the Wellcome Trust.
"It is essential that we develop our understanding of the underlying causes of the disease, and genome-wide association scans offer a powerful tool to do just this."