The American Society of Human Genetics sponsors an annual DNA Day Essay Contest. It's for grade 9-12 students from anywhere in the world.
This year's question is ...
Artificial intelligence (AI)—advanced computer systems that can quickly analyze large amounts of data—is being used in many areas of healthcare, from diagnosing diseases to recommending treatments. Now, experts are also using AI to help interpret genetic testing results, which examine your DNA to understand your risk for certain diseases or guide treatments. If you had to undergo genetic testing, would you want AI to be used to help make sense of your results?
It's a bit of a loaded question since it requires certain assumptions and seems designed to elicit positive responses to the current ability of AI to identify genetic diseases. This is reinforced by the request to identify the following points in the essay.
- When do you think using AI would be helpful in understanding your genetic test results?
- What risks or harms do you think using AI could pose in healthcare?
- What information would you want to learn from AI for analysis of your genetic information? Consider in comparison to the information provided in standard genetic test results.
Is this the best way to encourage high school students to be skeptical and practice critical thinking? Does anyone really believe that a typical high school student can understand what AI is and how if differs from simple data analysis? Does anyone really believe that they could evaluate the current and future role of AI in understanding genetic disease?
I suspect that what you will get is essays that simply regurgitate the standard hype in science publications and the media. Some students might even find select publications in the scientific literature that promote AI. I won't be surprised to discover that students accept AI predictions as definitive proof that the cause of disease has been discovered instead of realizing that this is only the first step and futher evidence from biochemistry and molecular biology is necessary.
I'd like to hear other opinions on whether this is a worthwhile effort. My own view is that it could be valuable if the questions were different; for example, you could ask students to discuss the definition of "gene" or current estimates of the human mutation rate. You could ask them to describe genetic disease that are caused by mutations in non-coding regions. You could ask a question to see if they understand the Neutral Theory of evolution and why there's so much genetic variation in the human population. Or you could even (gasp!) ask them how much of the human genome is junk.
These are questions that challenge students to understand fundamental principles of biology and get them to appreciate that there are controversies in the scientific literature. They also help focus attention on science and not on the applications of science (technology).
Here are the winners and their full essays.
11 comments :
Is that a question about genetics or about AI? (Seems more like the latter). I suppose it's within the organization's mandate to ask about the ethics of genetic testing, but the specific technology used to do the analysis seems orthogonal to that.
junk DNA is an old fashioned disputed theory.... Neutral theory was never on scientific consensus only a suggestion to answer Haldanes dilemma whitout any emprical evidence
The above comment was mine btw, forgot to log in.
@Anonymous I read Larry's book there are a lot of problem with that
So, Mehrshad, what proportion of the human genome could be deleted without consequence to the reproductive fitness of the individual concerned?
@Mehrshad Please explain why you think that junk DNA is an old fashioned theory.
And please explain why the Neutral Theory was never a scientific consensus. I'm sure there are some population geneticists who would would love to re-write their textbooks if you prove them wrong.
If there are "a lot" of problems with Larry's book you should have no difficulty describing three such problems.
Nothing all of it is required for proper genome functioning
@Larry _ and also there is no correlation between population effective size and genome size according to this recent paper :
Effective population size does not explain long-term variation in genome size and transposable element content in animals
Alba MarinoGautier DebaeckerAnna-Sophie Fiston-LavierAnnabelle HaudryBenoit Nabholz
ISEM, Univ Montpellier, CNRS, IRD, Montpellier, FranceUniversité Claude Bernard Lyon 1, LEHNA, UMR 5023, CNRS, Villeurbanne, FranceInstitut Universitaire de France (IUF), Paris, FranceUniversité Claude Bernard Lyon 1, LBBE, UMR 5558, CNRS, Villeurbanne, France
What is the reason for dismissing genetic load?
As to whether the neutral theory was a scientific consensus, wouldn't that depend on whether "the neutral theory" says that essentially all changes are neutral, or whether it says that there sure are lots of neutral changes? Contemporary population genetics does not say the former.
Post a Comment