Susan Murch, PhD, gave a talk at last week’s Dietary Supplement Analytical Summit put on by the United Natural Products Alliance in Salt Lake City. Murch, who is a chemistry professor at the University of British Columbia, said understanding the limitations of the technology is critical in its correct application. The recent hubbub about its use, first put into motion by the actions of New York Attorney General Eric Schneiderman, tends to overstate the reach of techniques that are currently available, she said.
Not all cells created the same
This is partly due to a misunderstanding about the differences between plant and animal cells. DNA taken from a cell is DNA, right? It is rarely that simple, Murch said. For one thing, in plant cells there are typically five sources of DNA, only three of which pertain to the plant species itself: DNA from the mitochondria, the DNA in the cell’s nucleus, and the DNA of the chloroplasts, the photosynthetic energy factories that occur only in plant cells, not in animal cells. Viruses and fungi are also commonly found in plant cells and can contribute their own DNA.
The process of using DNA to identify a species is reliant on choosing the right snippets, or primers, of the ultra long molecule that are characteristic of that particular species. Much of the DNA strand will be common to many other species. The primers are then amplified via a polymerase chain reaction, a technique now more than three decades old, to yield enough DNA to work with.
But not all primers are created equal. Some amplify better than others, and some differentiate between species more accurately than others. There are still many unknowns, Murch said.
“You find DNA that’s broken or sheared or has natural variations,” Murch said. “You are getting both false positives and false negatives.”
Murch cited a paper published this year in the journal Forensic Science International, Genetics. The paper’s authors put forth a possible combination of techniques to get around these conundrums, but noted: “While DNA barcoding enjoyed a remarkable success for animal identification . . . the attempts to identify a single barcode for plants remained a vain hope for a long time.”
Quality of sample
Where you take the DNA from within the cells matters, Murch said. DNA can vary from cell to cell and from location to location within the cell.
“The reason chloroplasts were selected by the international community is there is less variation in the chloroplasts than in the nucleus,” Murch said.
Another key limitation of DNA technology for botanical identification has to how raw materials typically come to a manufacturer. DNA is best taken from a raw plant material, and the fresher the better. DNA can be lost or degraded in almost every processing step, Murch said.
“Even the correct raw material can give false positives. In a dried sample, you are often dealing with DNA that is too degraded, too old, too lost,” she said.
Quality of process
In the end, Murch said, the judging quality of DNA as an identification strategy is no different than evaluating any other chemical test. Is it fit for purpose? Can it be reliably replicated?
“What is your sample, how is it prepared, how rugged are your methods? The degree of certainty for these DNA tests now used is between 54% and 85%. The public expects 100% certainty. These tests are not working as well as portrayed in the media, and none of the AGs are talking about this," she said.
“If you are only differentiating between species about 80% of the time the data is not reliable enough to launch lawsuits. And a failure to detect is not evidence of absence. If you tested Tylenol with a DNA test, you’d find corn. But you wouldn’t find Tylenol,” Murch said.