Executive Summary Continued:
DNAPrint Genomics - Company Advantages
Until now, it has been next to impossible to determine which genetic characteristics make a person a good responder or poor responder to a drug. This is because most drug reaction traits are multi-factorial and heterogeneous, and such problems are difficult to solve using hypothesis-driven research. As a result, scientists have focused on the "low hanging fruit" - simple traits caused by single genes. Unfortunately, 99% of human traits are complex not simple - the eye color of a child cannot be predicted very well simply by knowing the eye color of the parents. Geneticists have been using simple genetics methods for decades to try to understand the genetics of human iris color, and they have failed. More systematic mathematical methods for dissecting genetic pattern from genomics (as opposed to genetics) data are needed to solve these traits, yet they have not been introduced.
DNAPrint mined the human genome for a special type of SNP called an Ancestry Informative Marker (AIM), constructed genome maps of these AIMs and used them with novel algorithms to solve human iris color. The result is a validation of DNAPrint's novel method of using population structure to solve complex human traits, and the company's first product - RETINOME that will be used to give forensics investigators the iris color of an individual who has left DNA at a crime scene.
The challenges are the same whether one speaks of solving iris color, variable drug response (determining why it is that certain patients do not respond to certain drugs) or common diseases such as hypertension and diabetes. To solve these traits, it is necessary to perform pan-genome screening, but one cannot solve these traits simply by screening genomes with randomly selected or publicly available markers. Our competitors do this - they select a small number of SNPs in a small number of genes (called a "Candidate gene" approach) in an attempt to study extreme responder and non-responder phenotypes. If they pick the wrong SNPs or genes, they do not find a solution - this is the problem with the hypothesis-driven approach. Genome based linkage analysis in large populations of unrelated individuals is not cost effective because it costs several hundreds of thousands of dollars per patient to do this type of study properly. Some companies use "isolated populations" to bring the costs down, but this approach carries significant disadvantages because their results do not always generalize to the world population. The result of all of these problems is that no accurate, powerfully predictive genomics based tests have yet been introduced to market.
DNAPrint's will be the first because we are probably one of a handful of scientists in the world that understand how to use something called "population structure" to solve complex human traits. DNAPrint uses an original approach called Mapping by Admixture Linkage Disequilibrium (MALD) with proprietary genomic maps and algorithms (collectively called the ADMIXMAP platform). The scientific advantage is that the markers that constitute our maps are precious ones in that they carry an unusual amount of information on population structure - that is to say, one can infer something about population affiliation (whether continental, intracontinental, or sub regional) by reading each marker. We do not usually seek to identify the actual genes causing the trait (unless we need to for a special reason), but instead we seek to identify those AIMs that are correlated with the genes in a statistical sense. Most human traits are inherited as a function of population affiliation on one level or another - blue iris color is unique to Indo Europeans, prostate cancer is of greater incidence in African Americans, diabetes in Native Americans, and virtually every drug shows different response characteristics in different human populations.
DNAPrint's secret is that exquisitely accurate classification tests can be developed for these problems simply by measuring AIMs! The economic advantage of the platform is that we can screen a genome in 100 individuals for about 1% of the cost of standard methods in existence today (because LD is extreme in recently admixed populations). The practical advantage is that we can develop classification tests without having to identify the actual genes involved. The scientific advantage is that we could easily identify those genes by increasing the marker density within each relevant region.
Using ADMIXMAP, we do better, more systematic and objective science, but we also guarantee that any solutions we find are constructed of private, not public, markers that have ownership implications. Competitors who rely on the public database not only struggle with statistical power, but they will be fighting over the same "SNPs" in the patent office.
Our main advantage is our math. DNAPrint Genomics Inc. was founded by a team of scientists with research and commercial experience in high-level mathematical and statistical modeling, programming and molecular genetics. Mathematical / Statistical genetics talent is limiting and math workflow is a bottleneck faced by our competition. Sound and original math represents an excellent barrier to entry.
We have focused on building one of private industries best analytical teams within the field of population genomics, have begun patenting innovative math tools for this type of research and we have already filed patent for five different mathematical methods and software algorithms useful for finding complex genetic pattern in high-density genomics data sets. These methods have been validated through their performance in the development of the products described in this plan.
The flagship DNAPrint patent is that which describes the 8,000 AIMs and statistical methods needed to perform Admixture Screening using SNPs. We believe that someday those in the field will look back on this patent as a landmark patent because we are the first to realize that most human traits can be understood through an exquisitely accurate measurement of human population structure and sub-structure. Why is it that redheads require 20% more anesthesia than other individuals of European descent? Why do they exhibit excessive bleeding and hypertension under certain anesthetics? Certainly not because of a gene for red-hair! Rather, the complex milieu of genes that predispose to these responses is correlated with Northwestern European biogeographical ancestry - Irish, British, and some Scandinavian. If one considers all of humanity in one giant family tree, and indo-Europeans as a branch of that tree, and Northwest European Anglos as a sub-branch, you would see that humans with red hair constitute a very large extended family, and the aberrant responses to anesthesia "runs in the (extended) family". The markers in the genome necessary for predicting the response are a subset of those that allow you to measure this population structure - to partition this extended family from others. This is the basis of our methodology, and it is extremely powerful, but it is also a new idea, which promises to be controversial even within the genetics community, though we have a large contingent of world-renowned population geneticists that not only support our approach but also are fascinated by it! Our early results speak for themselves!
In addition to this strategic advantage, there are practical advantages of our method. Screening genomes with today's technology is far too expensive, yet most human traits are sufficiently complex that pan-genome screening is required in order to understand them. Multiple genes interacting with each other determine physical characteristics such as height and weight. So too are drug responses or disease susceptibility (i.e. cancer). There is no drug interaction gene or cancer gene; there are several 10s or even 100s of genes that determine these traits. It is well known that over 99% of known phenotypes (physical traits) are more a function of complex multi-gene interactions within and between genetic pathways that single genes.
Most scientists agree that complex genetic analysis will be REQUIRED for the development of personalized medical products, such as those we are developing to match patients with drugs most appropriate for their genetic condition. Those companies with the math that enable this type of analysis will hold a significant advantage.
For example, tests developed by some of our competitors tend to rely on inferences from only one gene, or tend to rely on analysis performed at the level of the SNP (as opposed to the level of complex sets of SNPs called haplotypes). Such tests will be too simplistic to enable sensitive or accurate inferences to be drawn in the clinic. Few companies, including Celera, Motorola, or Human Genome Sciences, for example, have yet presented complex genetics analytical tools or "solutions" and we are one of a handful of companies to focus on math as an enabling tool for the development of products for the emerging personalized medicine markets.
The use of Ancestry Informative Markers (AIMs) with Admixture Mapping represents the best, most economical means by which to solve complex human traits and develop classification (predictive diagnostics) tests.
