I attended the Lab Tour session this morning via live stream. I tried to attended the Advanced Autosomal session at 5:00, but I had lots of problems with the feed.
The Lab Tour session was very interesting and included a demonstration on how DNA is extracted using a strawberry. The presenter then explained a little about the different tests available. Volunteers carrying colored circles were enlisted to demonstrate how mtDNA results are produced. Autosomal DNA lab analysis was explained in the same way, but this time the volunteers carried large cards with colored dots representing SNPs. Apparently the cards weren't correctly printed, so there was a lab failure as an audience member put it.
I didn't know that the Y DNA and mtDNA samples weren't analyzed using a chip? Only the autosomal testing is done with a chip. That's probably why the autosomal tests come back so quickly.
All of us are wishing for more accurate ethnicity results; but, as the presenter explained we are a long way from that goal at this time. As she stated all the DNA companies use different reference populations; which is the reason all the companies produce different results. For instance Family Tree DNA uses these data sets GeneByGene DNA customer database, Human Genome Diversity Project, International HapMap Project, Estonian Biocentre data. 23andme uses the Human Genome Diversity Project, International HapMap Project, and 1000 Genome Project. Ancestry uses the Human Genome Diversity Project and the Sorenson database. All of the companies use their own customer sample collections. So we don't get consistent results.
To explain why it's so difficult to accurately predict ethnic origins the presenter used a US Map game. She asked which states are associated with football? Many states are. So we can't pinpoint one. If we have some additional information like a Football state and the Fighting Irish we can confidently pinpoint the state. Not being an expert on the subject I interpret this as meaning we need to find certain combinations of SNPs in order to establish a link to a particular place. I also suppose we don't have all of these combinations mapped out yet. We need larger sets of data from every ethnic population in the world in order to find the unique SNP combinations shared by these groups. That may be a wrong interpretation? It's the way I understood it (I shouldn't have napped in biology class). We are still in the infancy of ethnic prediction. The data is still being collected, and the analysis process is still being refined. From this explanation it seems like eventually we will get the accurate results we are looking for.
The presenter also shared a story about when she didn't follow lab procedures and temporarily blinded herself. That was in the early days of analysis when goggles were required so your eyes wouldn't be damaged by ultra violet light. The process is safer now. DNA and the Genealogical proof standard will be live streamed at 2:30 pacific time tomorrow. It's a free session.