If you had to undergo genetic testing, would you want AI to help analyse your results?

Anaiya (Year 12)

Editor’s note: Year 12 student Anaiya entered this fascinating essay into the American Society of Human Genetics (ASHG) Annual DNA Day Essay Competition. DNA Day commemorates the completion of the Human Genome Project in April 2003 and the discovery of the double helix of DNA in 1953. This contest is open to students in Year 10-13 worldwide and asks students to examine, question, and reflect on important concepts in genetics. Essays are expected to be well-reasoned arguments that indicate a deep understanding of scientific concepts related to the essay question. They are evaluated by ASHG members through three rounds of scoring. CPD

Recently, artificial intelligence has become considered to be a powerful tool in data analysis with several positive impacts which could be applied to genetic testing, allowing us to gain more insights into hereditary diseases, cancers, or rare genetic disorders on a new level of detail. In this essay, I will be discussing the benefits and drawbacks of using AI in genetic testing, and whether I would accept the involvement of AI if I were to undergo genetic testing.

When AI would be helpful in understanding genetic test results:

A main benefit of AI is its level of accuracy when analysing large amounts of data. In the case of genetic testing, accuracy is crucial as even the smallest mutation found somewhere in a person’s genome can have a major effect on how they function. AI algorithms could be used to detect chromosomal variants that can easily be missed by the human eye or existing software such as FISH (7), helping patients and doctors to make more informed decisions from results of genetic tests.

Given the fact that AI can filter through such large datasets, whole genome sequencing (WGS) (2) could become a genetic test that is carried out more often. It is not commonly practiced as it is extremely time consuming, and only the exome contains useful information. However, AI can speed up this process drastically as it is able to interpret large amounts of data much faster but still with incredible detail, allowing us to find more precise information than with any other genetic test, such as karyotype testing (3) . As the whole genome is analysed, inherited diseases, cancerous mutations, or specific variants could be recognised.

It is widely acknowledged that AI is mainly used because it can carry out tasks quickly but still to an impressive standard. Therefore, it can also help to reduce costs and increase efficiency in our healthcare systems because more tests can be processed as there won’t be the need for as many healthcare professionals at a time, resulting in less wasted resources. This is beneficial for patients as they will have reduced waiting times, and there will be more time to take proactive steps to address any issues that the tests show.

Risks or harms using AI could pose in healthcare:

On the other hand, AI is a complicated and seems daunting to many which could be a problem in the healthcare industry as many people could be sceptical about AI and would rather their genetic tests were analysed manually instead. Therefore, it might take time for artificial intelligence to be incorporated into the world of healthcare due to the negative connotations behind it.

Furthermore, a common fear regarding AI is its ability to take over jobs. Possibly, if AI becomes too overused and relied on by clinicians, jobs may be replaced as AI is quicker and generates accurate information. However, if more trustworthy doctors are losing their jobs and AI is relied on too much, it might make mistakes without healthcare professionals being present to oversee AI’s activity.

Information AI could provide for analysis of genetic information in comparison to the information provided in standard genetic test results:

Currently, some commonly used genetic tests are microarray which investigates specific genes within chromosomes, and karyotype testing which focuses on the entire set of chromosomes. I would want AI to incorporate both the higher resolution and zoom in microarray to observe aneuploidies (4) with the aspect of karyotype testing where structural rearrangements (5) that can reveal problems with infertility or cancer can be observed.
Also, prenatal genetic tests such as CVS (6) and amniocentesis are carried out very often in the first and second trimester to detect abnormalities, defects, or genetic diseases in the foetus (7). From AI, we could learn information on the possibility of NTDs (neural tube defects) because CVS does not provide any. AI could be utilized to immediately find correlations and predict if any NTDs will occur in the future so that other follow up tests will not be necessary, reducing discomfort for patients.

Would I want AI to be involved in evaluating a genetic test of my own? In conclusion, from the presented evidence, I would prefer for AI to be involved in evaluating my results alongside healthcare professionals because I would feel satisfied with the accuracy of my results and the suggested treatment options personalised for me.

References

1. Dutra, Amalia. Fluorescence In Situ Hybridization (FISH). National Human Genome Research Institute. [Online] 1st March 2025. [Cited: 28th February 2025.] https://www.genome.gov/genetics-glossary/Fluorescence-In-Situ-Hybridization-FISH.
2. Francois Balloux 1, Ola Brønstad Brynildsrud, Lucy van Dorp, Liam P Shaw, Hongbin Chen, Kathryn A Harris, Hui Wang, Vegard Eldholm. From Theory to Practice: Translating Whole-Genome Sequencing (WGS) into the Clinic. PubMed – National Library of Medicine. [Online] 4th September 2018. [Cited: 1st March 2025.] https://pubmed.ncbi.nlm.nih.gov/30193960/.
3. Uma M. Reddy, M.D., M.P.H., Grier P. Page, Ph.D., George R. Saade, M.D., Robert M. Silver, M.D., Vanessa R. Thorsten, M.P.H., Corette B. Parker, Dr.P.H., Halit Pinar, M.D., Marian Willinger, Ph.D., Barbara J. Stoll, M.D., Josefine Heim-Hall, M.D., Michael. Karyotype versus Microarray Testing for Genetic Abnormalities after Stillbirth. New England Journal of Medicine. [Online] 6th December 2012. [Cited: 1st March 2025.] https://www.nejm.org/doi/full/10.1056/NEJMoa1201569.
4. Eric Graham. Karyotype Testing Explained. Cofertility. [Online] 29th January 2025. [Cited: 20th February 2025.] https://www.cofertility.com/family-learn/karyotype-testing.
5. Karyotype. National Genomics Education Programme. [Online] 17th August 2022. [Cited: 21st February 2025.] https://www.genomicseducation.hee.nhs.uk/genotes/knowledge-hub/karyotype/#advantages-and-limitations-of-karyotyping.
6. Unknown. Chorionic Villus Sampling (CVS). Johns Hopkins Medicine. [Online] unknown unknown unknown. [Cited: 24th February 2025.] https://www.hopkinsmedicine.org/health/treatment-tests-and-therapies/chorionic-villus-sampling-cvs.
7. Zarko Alfirevic, Kate Navaratnam, Faris Mujezinovic. Amniocentesis and chorionic villus sampling for prenatal diagnosis. PubMed – National Library of Medicine. [Online] 4th September 2017. [Cited: 1st March 2025.] https://pubmed.ncbi.nlm.nih.gov/28869276/.