Paper describing the development and function of the AlzGene database of genetic studies in Nature Genetics.  Project is first of its kind.

The AlzGene database is a revolutionary development that provides an unbiased and regularly updated collection of genetic association studies performed on Alzheimer’s disease phenotypes. This site allows researchers to post their findings and comment on other findings, and for the first time, aggregates huge amounts of information that will help in the search for the genetic makeup of Alzheimer’s disease. Currently, the database contains more than 1,000 studies. The site also has a continuously updated list displaying the genes most strongly associated with AD (on the website see “Top AlzGene Results”).

Led by Dr. Lars Bertram, Assistant Professor of Neurology at Harvard Medical School, a team of researchers at Massachusetts General Hospital has developed the database in collaboration with the Alzheimer’s Research Forum. Dr. Bertram and his colleagues have just had their paper describing the development and function of Alzgene published in Nature Genetics, a leading journal in the field (abstract available).

Cure Alzheimer’s Fund is proud to be supporting this effort with complete annual funding for Alzgene, and congratulates Dr. Bertram and his colleagues for this achievement and significant contribution to the fight against Alzheimer’s disease. AlzGene is the first project of its kind for any of the genetically complex diseases.

To give you a full understanding of Alzgene, we wanted to provide you with a description of the site by its founder, Dr. Bertram. Some of the language is very technical, so we’ve included a basic genetic vocabulary primer to help you understand his description. In addition, as you read more about the role of genetics in a variety of diseases, as well the increasing understanding of how genetics affects our normal, healthy lives, it may be useful to become familiar with some key concepts:

Allele: any one of the alternative forms of a given gene. For example, the APOE gene has four major alleles, or forms, the APOE4 being the one with the most firmly established risk factor for AD.

Penetrant: strength of effect. A highly penetrant gene or gene allele may not appear very frequently in the population, but when it does it has a strong effect.

Prevalence: frequency of occurrence. If an allele appears in a high percentage of the population, it is highly prevalent.

Autosomal dominant: an allele that masks the expression of another allele — a strong genetic characteristic.

Mendelion inheritance: one method in which genetic traits are passed from parents to offspring.

Polymorphism: defined in context below.

Meta-analysis: defined in context below.

Dr. Bertram describes the Alzgene project as follows:

The project aims to bring clarity to the increasingly prolific, but no less obscure field of late-onset Alzheimer’s disease (AD) genetics which is currently fueled by nearly a dozen new findings each month.

Genetically, AD is complex and heterogeneous and appears to
• follow an age-related dichotomy in which
• rare and highly penetrant early-onset familial AD mutations are transmitted in an autosomal dominant fashion,
• while risk for late onset AD without obvious Mendelian inheritance is believed to be modulated by genetic variants with relatively low penetrance, but high prevalence.

To date, the only firmly established genetic risk factor for AD is the ε4-allele of APOE. Beyond this, nearly 1,000 papers have been published claiming or refuting association between AD and literally hundreds of putative risk alleles in other genes, but none of these findings has yet received a comparable degree of independent replication as APOE.

For AlzGene, my colleagues and I have systematically collected genotype and key demographic data of these ~1,000 studies and summarized them on a publicly available database. In addition to displaying an exhaustive collection of all genetic association studies published in the field of AD, genetic markers (a.k.a. “polymorphisms”, representing particularly common forms of genetic mutations) with published genotype data available in three or more independent samples have been subjected to systematic meta-analyses, a statistical technique which provides a quantitative and essentially unbiased summary of any marker’s chance of being a bona-fide AD risk factor when considering all of the published evidence at once.

The vast majority of these analyses show no significant summary effects, suggesting that the tested polymorphisms probably do not represent major AD risk factors. On the other hand, a few variants show significant risk (or protective) effects when the data across all published studies is meta-analyzed. With the exception of APOE-ε4, the associated odds ratios (which signify by how much an individual’s risk to develop AD may be increased or decreased if he/she is a carrier of such a polymorphism) are small, yielding increases in disease risk between ~1.1 and 1.4-fold. This compares to an increase between ~3 and 15-fold for carriers of one or two copies of the APOE ε4-allele, respectively. Despite their small effect size, the genetic variants uncovered in the AlzGene meta-analyses may still be important on a population-wide level owing to their relatively high frequency in the general population. A “typical” risk-allele may be present in 20% or more of all people, and – together with other genetic and non-genetic risk factors – may significantly contribute to an individual’s risk of developing AD.

Among these minor, but from a statistical standpoint promising findings are genes such as PSEN1 (encoding presenilin 1, mutations in which are the leading cause of early-onset familial AD), TF (transferrin, involved in iron homeostasis), PRNP (prion protein, which is known to affect the risk of developing transmissible spongiform encephalopathies), CST3 (cystatin C, which also causes a related hereditary neurodegenerative disease), and MAPT (encoding tau, one of the pathological hallmarks in brains of deceased AD patients).

Two of the newest additions to the list are SORCS1 (originally identified by Grupe and colleagues in 2006), and SORL1 (published by Rogaeva et al. in the February 2007 issue of Nature Genetics), which may both be involved in the in a common pathway regulating the sorting and processing of APP.

As is the case for these and all other proclaimed novel AD risk genes, substantially more data from independent research groups are needed in order to judge whether the currently promising meta-analysis findings will hold up over time and support a genuine disease-modifying role of these variants.