Phil Rogers, a Chicago news reporter, recently sent samples from home test kits to several online DNA sequencing services.
One told him his ancestors were from Ireland and Scotland. Another said Portugal. Others said Scandinavia, Peru, and Afghanistan.
How can this be?
There are three components of computing:
- Data is information that has been turned into numbers.
- Databases are collections of data organized for rapid search and retrieval.
- Algorithms are formulas devised by persons for solving problems.
Each company he mailed his DNA to used different information, different data sets, and different formulas. And, of course, each claimed to be superior to the others.
When information, databases, and algorithms vary, results vary, and usually quite widely.
UCLA geneticist Dr Wayne Grody says there might be some vague truths somewhere, but home DNA test kits are mostly like the quack cure-alls pitched by snake oil salesmen a hundred years ago.
Artificial Intelligence isn’t artificial at all, but an unfortunately chosen term used to describe Machine Learning.
Most people think A.I. is something new, even though it’s been around since 1956, when scientists began to develop “thinking machines.” Encyclopedia Britannica says A.I. is “the ability of a computer to perform tasks commonly associated with intelligent beings.” Think of it as the ability to recognize patterns and perform calculations within huge data sets at blindingly fast speeds.
But as many have demonstrated, A.I.’s information and databases vary greatly, and algorithms ignore context, which leads to many completely meaningless conclusions. Click here to see the 3-minute article on Big Data – and the bonus Marvin Gaye YouTube video – from January 2017.
The reporter submitted DNA from his Labrador Retriever to the same services. One company did not notice the sample was non-human, and concluded in their report that this person would be great at basketball. The first thing that came to my mind was Charles Barkley.