According to Gram Research analysis, dolphins that learn to use sponges as tools have distinctly different DNA methylation patterns—chemical tags on their genes—compared to dolphins that don’t use tools. Scientists studied 96 wild dolphins and found that artificial intelligence could identify tool-users based on their DNA patterns 68% of the time, suggesting that learned behaviors can leave measurable marks on our genetic code at the molecular level.
Scientists discovered something remarkable: dolphins that use sponges as tools to find food have different DNA patterns than dolphins that don’t use tools. Researchers studied 96 Indo-Pacific bottlenose dolphins in Australia and analyzed over 29,000 spots in their DNA where chemical tags could attach. Using artificial intelligence, they found that these chemical tags could predict whether a dolphin was a tool-user about 68% of the time. This suggests that learned behaviors—skills passed down from parent to child—can actually change how our genes work at a molecular level, a discovery that could help us understand how culture shapes biology in humans and animals alike.
Key Statistics
A 2026 study of 96 Indo-Pacific bottlenose dolphins in Shark Bay, Australia found that machine learning models could identify tool-using dolphins based on DNA methylation patterns with 68% accuracy, significantly better than random chance.
Researchers analyzed over 29,812 specific sites in dolphin DNA where chemical tags attach, finding that tool-using dolphins showed distinct methylation patterns compared to non-tool-users in the same population.
Tool-using dolphins in the study exhibited different diets, social networks, and survival rates compared to dolphins in the same habitat that lacked the culturally transmitted sponge-foraging technique.
The 2026 research involved 23 sponge-using dolphins and 73 non-spongers, demonstrating that a learned behavior passed through dolphin culture correlates with measurable changes in how genes are chemically regulated.
The Quick Take
- What they studied: Whether dolphins that learn to use sponges as tools have different chemical patterns in their DNA compared to dolphins that don’t use tools
- Who participated: 96 Indo-Pacific bottlenose dolphins living in Shark Bay, Western Australia—23 that use sponges as foraging tools and 73 that don’t
- Key finding: DNA methylation patterns (chemical tags on genes) could correctly identify tool-using dolphins 68% of the time, suggesting learned behavior leaves a measurable mark on our genetic code
- What it means for you: This research shows that behaviors we learn and pass down through culture can actually change how our genes work. While this study focused on dolphins, it opens doors to understanding how human culture and learning might similarly influence our biology at the molecular level
The Research Details
Scientists collected skin samples from 96 wild dolphins in Shark Bay, Australia. Some dolphins in this population have learned to use marine sponges as tools to protect their noses while searching the seafloor for fish—a skill taught by mothers to calves. Other dolphins in the exact same area never learned this trick. The researchers analyzed over 29,000 specific locations in the dolphins’ DNA where chemical tags (called methyl groups) can attach. These chemical tags don’t change the DNA sequence itself, but they act like switches that turn genes on or off. This process is called DNA methylation, and it’s one way that environment and experience can influence how genes work.
The team used machine learning—a type of artificial intelligence—to see if they could predict which dolphins were tool-users just by looking at their methylation patterns. The computer was trained on the DNA patterns from some dolphins and then tested on others it had never seen before. This approach helps ensure the results aren’t just due to chance.
The dolphins were also studied for their diet, social connections, and survival rates. Tool-using dolphins had different diets and social networks compared to non-tool-users, suggesting that this learned behavior genuinely changes how they live.
This research is important because it bridges two big ideas in biology: culture and genetics. For a long time, scientists thought these were completely separate. Culture was about learned behaviors, while genetics was about DNA. This study suggests they’re actually connected—that learning and culture can leave physical marks on our genes. This matters for understanding how humans and other animals adapt to their environments and pass knowledge to future generations.
The study was published in a highly respected scientific journal (Philosophical Transactions of the Royal Society B), which means it went through rigorous peer review. The researchers used a large sample of wild dolphins and analyzed thousands of genetic locations, making their findings more reliable. The machine learning model performed significantly better than random chance, which is important. However, the model’s accuracy of 68% means it’s not perfect—there’s still a lot we don’t understand about how behavior and DNA methylation are connected. The study was observational rather than experimental, so we can’t say for certain that the tool-use behavior caused the DNA changes, only that they’re associated.
What the Results Show
The machine learning model could correctly identify whether a dolphin was a tool-user based on DNA methylation patterns about 68% of the time. This is better than random guessing (which would be 50%), and statistical tests confirmed this wasn’t just luck. The researchers found that specific locations in the dolphins’ DNA had different methylation patterns between tool-users and non-tool-users.
Tool-using dolphins showed distinct patterns at 29,812 different CpG sites (these are specific spots in DNA where methylation commonly occurs). While the model couldn’t perfectly predict behavior from DNA alone, the fact that it performed significantly better than chance suggests that DNA methylation patterns genuinely contain information related to sponging behavior.
The dolphins that used sponges as tools also had measurably different diets, social networks, and survival outcomes compared to non-tool-users. This indicates that the learned behavior isn’t just a quirk—it genuinely changes how these dolphins live their lives.
The research revealed that cultural behavior (tool use) correlates with differences in social connections and diet. Tool-using dolphins associated with other tool-users and had access to different food sources than non-tool-users. These lifestyle differences may explain some of the DNA methylation changes observed. The study also demonstrates that epigenetic changes—modifications to how genes are expressed—can be associated with culturally transmitted behaviors, not just with genetics or random chance.
This is one of the first studies to directly connect culturally learned behavior with epigenetic changes in wild animals. Previous research has shown that DNA methylation can change based on environment, stress, and diet, but linking it to specific learned behaviors passed through culture is novel. The findings align with growing evidence that epigenetics plays a role in how organisms adapt to their environments and that these changes can sometimes be passed to offspring, though this study didn’t examine inheritance.
The study’s main limitation is that it’s observational—researchers watched dolphins and measured their DNA, but couldn’t control variables or run experiments. This means we can’t definitively say that learning to use sponges causes the DNA changes; they’re just associated. The machine learning model’s 68% accuracy, while better than chance, means it’s not a reliable predictor on its own. The study also only looked at one population of dolphins in one location, so results might not apply to other dolphin populations. Additionally, the researchers measured DNA methylation at one point in time, so they couldn’t track how these patterns change over a dolphin’s lifetime or whether they’re passed to offspring.
The Bottom Line
This research is primarily important for scientists studying how behavior and biology interact. For the general public, it suggests that learned skills and cultural practices may have deeper biological roots than previously thought. While we can’t yet make specific health or lifestyle recommendations based on this dolphin research, it opens the door to future studies examining whether human learning and cultural practices similarly influence our epigenetics. Confidence level: This is early-stage research that needs follow-up studies before making practical recommendations.
Scientists studying evolution, behavior, genetics, and epigenetics should pay close attention to this work. Researchers interested in how culture shapes biology in humans will find this dolphin study valuable. Educators and those interested in animal cognition may also find this research fascinating. This research is not directly applicable to individual health decisions at this time.
This is fundamental research exploring how behavior and genetics connect. Practical applications—if any—would likely take many years to develop. The immediate impact will be on scientific understanding and future research directions rather than on individual behavior or health.
Frequently Asked Questions
Can learned behaviors actually change your DNA?
Learned behaviors don’t change your DNA sequence itself, but they can change DNA methylation—chemical tags that turn genes on or off. This dolphin study shows tool-use behavior correlates with different methylation patterns, suggesting culture and biology are more connected than previously thought.
How accurate is DNA methylation at predicting dolphin tool use?
Machine learning models using DNA methylation patterns correctly identified tool-using dolphins about 68% of the time—better than random guessing but far from perfect. This suggests methylation contains some information about behavior, but many other factors also influence whether dolphins use tools.
Do these DNA changes get passed to baby dolphins?
This study didn’t examine whether methylation changes are inherited by offspring. Researchers only measured DNA patterns at one point in time. Future studies would need to track dolphins across generations to answer whether these epigenetic changes pass to calves.
Could this research apply to humans learning new skills?
This dolphin research suggests learned behaviors may influence epigenetics in animals, but we don’t yet know if human learning works the same way. More research is needed before we can say whether human skills and culture similarly change our DNA methylation patterns.
Why does it matter that culture and DNA are connected?
Understanding how learned behaviors influence our genes at the molecular level helps explain how culture shapes biology. This bridges two fields—genetics and anthropology—and could eventually help us understand human evolution, adaptation, and how knowledge passes between generations.
Want to Apply This Research?
- While this research doesn’t directly apply to personal health tracking, users interested in learning and skill development could track new skills learned and note any changes in energy, mood, or social connections over time. This could help individuals observe whether learning new behaviors correlates with other life changes.
- Users could use this research as motivation to learn new skills or cultural practices, understanding that these learned behaviors may have deeper biological significance than previously appreciated. The app could encourage users to document skills they’re learning and reflect on how these new behaviors change their daily lives and social connections.
- Over months or years, users could track the development of new skills or cultural practices and note correlations with changes in lifestyle, diet, social networks, or wellbeing. This long-term perspective mirrors the dolphins’ experience of learning and integrating new behaviors into their lives.
This research describes associations between learned behavior and DNA methylation patterns in dolphins and does not provide medical advice for humans. The study is observational and cannot prove that behavior causes epigenetic changes. While fascinating for understanding the biology of culture, these findings should not be interpreted as direct guidance for human health or behavior modification. Individuals with questions about genetics, epigenetics, or health should consult qualified healthcare providers or genetic counselors. This article is for educational purposes and should not replace professional medical or scientific consultation.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.