Non-genetic inheritance of stochastically induced behavioral individuality in a naturally clonal fish

This study provides the first evidence that stochastically induced behavioral differences in naturally clonal fish can be non-genetically transmitted across generations, as mothers with higher feeding activity produce more active offspring, suggesting such variation may significantly contribute to evolutionary processes and population adaptability.

The Gist

The Big Idea: Can "Random" Habits Be Inherited?

Imagine you have a factory that makes identical twins. They have the exact same DNA (blueprint), they grow up in the exact same room, eat the exact same food, and sleep in the exact same beds. By all logic, they should act exactly the same way, right?

Usually, scientists thought that if you remove all genetic differences and all environmental differences, the only thing left is "noise"—random, meaningless glitches. Like static on an old TV.

This paper says: "No, that static is actually a signal."

The researchers studied a special fish called the Amazon Molly. These fish are clones; they are genetic copies of their mothers. The scientists put 34 of these "mother" fish and their 232 "baby" fish into perfectly identical, sterile environments. No differences in water, no differences in food, no differences in genetics.

The Discovery: The "Personality" of the Clone

Even though these fish were clones in a clone-friendly room, they didn't act like robots. They developed distinct personalities.

• Some fish were like hyperactive toddlers, swimming around constantly.
• Some were laid-back couch potatoes, barely moving.
• Some were foodies, spending hours staring at the food patch.
• Some were fast eaters, grabbing a bite and zooming off.

The scientists tracked them for thousands of hours (about 19,000 hours of video!) and found that these differences weren't just random flukes. They were consistent. A fish that was a "foodie" one day was likely a "foodie" the next.

The Twist: The "Grandma Effect" (But with Moms)

Here is where it gets really weird and fascinating. The scientists asked: "If Mom is a certain way, will her babies be the same way?"

Since the babies are clones and grew up in the same room as the moms, there shouldn't be any reason for them to be similar. But they found a surprising link:

• The Activity Link: If a mom was super active, her babies were not necessarily active.
• The Feeding Link: If a mom spent a lot of time eating, her babies turned out to be super active.

The Analogy: Imagine a mother who is a "foodie." She spends all her time at the buffet. You might expect her kids to be foodies too. Instead, her kids turn out to be marathon runners! It's a cross-trait inheritance. The mom's habit of eating somehow "programmed" the baby's habit of moving.

The study found that the mom's eating habits explained about 33% of why the babies were active. That is a huge amount of influence for something that isn't DNA!

Why Didn't Size Matter?

The researchers wondered: "Maybe the moms who ate more just got bigger, and big moms have big, active babies?"

They checked. Nope.

• Eating more didn't make the moms bigger.
• Being a bigger mom didn't make the babies bigger.
• The babies' size didn't explain why they were active.

So, it wasn't about physical size. It was about something invisible.

The "Invisible Backpack" Analogy

Think of the mother fish as carrying an invisible backpack of biological information.

• When she spends time eating, she isn't just filling her stomach; she is changing her internal chemistry (hormones, energy levels, or perhaps "epigenetic" switches).
• She passes this invisible backpack to her babies when they are born.
• The babies don't inherit the behavior (eating) directly. Instead, they inherit the state (a specific chemical mix) that makes them want to run around (be active).

It's like a mother who drinks a lot of coffee. She doesn't pass the coffee to her baby, but she passes on a "caffeinated state" that makes the baby jittery and energetic.

Why Does This Matter?

For a long time, scientists thought that if you didn't have genetic mutations (changes in DNA), you couldn't evolve. They thought "stochastic" (random) variation was just useless noise.

This paper suggests that random noise can actually be useful.

• Even without new DNA, a population can change its behavior from one generation to the next based on these invisible "backpacks."
• This gives clonal species (like the Amazon Molly) a secret superpower: they can adapt to changes in the world much faster than we thought, without waiting for slow genetic mutations.

The Bottom Line

Nature is messy. Even when you try to make everything perfectly identical, life finds a way to create differences. And surprisingly, those random differences in a mother's daily habits can be passed down to her children, shaping their future in ways we are only just beginning to understand.

In short: You don't need a new DNA blueprint to change your family's future; sometimes, just changing your daily routine is enough to leave a mark on the next generation.

Technical Summary

1. Problem Statement and Research Gap

• Context: Biological research has increasingly recognized "stochastic phenotypic variation"—differences in traits (behavior, physiology, gene expression) that arise despite the apparent absence of genetic and environmental differences. In naturally clonal organisms (like the Amazon molly, Poecilia formosa), substantial behavioral individuality (e.g., in activity and feeding) emerges even under highly standardized conditions.
• The Gap: While the existence and fitness relevance of this stochastic variation are established, it remains unknown whether such variation is transmitted across generations. If stochastically induced differences can be inherited non-genetically, they could constitute a previously unrecognized mechanism driving evolutionary processes and population adaptability.
• Hypothesis: The authors hypothesize that stochastic behavioral differences in mothers (specifically activity and feeding) predict similar differences in offspring, potentially mediated by maternal state (e.g., body size, physiological condition) rather than genetics or shared environment.

2. Methodology

The study employed a rigorous two-generation experimental design using the naturally clonal Amazon molly (Poecilia formosa).

• Study Subjects:
  • Generation 1 (Mothers): $N=34$ genetically identical females.
  • Generation 2 (Offspring): $N=232$ offspring from the 34 mothers.
  • Genetic Control: All mothers were clones derived from a single lineage; offspring were produced via gynogenesis (sperm from a related species triggers development but contributes no DNA), ensuring genetic identity within the maternal line.
• Environmental Control:
  • Mothers were separated immediately after birth into near-identical individual tanks.
  • Conditions were strictly standardized: temperature ($24 \pm 1^\circ$C), light cycle (12:12), and diet.
  • Behavioral Typing (Mothers): Recorded for 10 hours/day over 28 days.
  • Behavioral Typing (Offspring): Recorded for the first week of life (mean $\approx$ 5.1 days) using the same protocol.
• Data Collection:
  • High-Resolution Tracking: Video recorded at 5 frames/second using Basler cameras and tracked via Biotracker software.
  • Metrics:
    • Activity: Average swimming speed (cm/sec) during 8 hours of no food.
    • Feeding: Time spent (minutes) in a defined "feeding zone" during a 2-hour feeding window.
  • Scale: The study generated $\sim$337 million data points over $\sim$18,740 observation hours.
• Statistical Analysis:
  • Repeatability ($R$): Calculated using Linear Mixed-Effects Models (LMMs) to quantify among-individual variance vs. within-individual variance.
  • Transgenerational Transmission: Linear models tested if maternal behavior predicted offspring behavior.
  • Mediation Analysis: Tested whether the link was mediated by maternal size at parturition or offspring size at birth.
  • Data Aggregation: Analyses were performed at the brood level (averaging offspring per mother) to estimate narrow-sense heritability analogues, and at the individual level (repeated measures) to test robustness.

3. Key Results

A. Emergence of Stochastic Individuality

• Mothers: Significant among-individual repeatability was found for both activity ($R = 0.356$) and feeding time ($R = 0.201$), despite genetic and environmental uniformity.
• Offspring: Similarly, offspring exhibited significant individuality in activity ($R = 0.295$) and feeding ($R = 0.203$).
• Conclusion: Stochastic behavioral individuality is a robust phenomenon in both generations.

B. Transgenerational Transmission (The Core Finding)

• Cross-Trait Prediction: Maternal feeding behavior significantly predicted offspring activity.
  • Mothers that spent more time feeding produced more active offspring.
  • Effect Size: Maternal feeding explained ~33% ($R^2 \approx 0.329$) of the total variation in offspring activity ($p = 0.001$).
• No Within-Trait Transmission:
  • Maternal activity did not predict offspring activity ($p = 0.089$, trend only).
  • Maternal feeding did not predict offspring feeding ($p = 0.628$).
  • Maternal activity did not predict offspring feeding ($p = 0.740$).
• Robustness: The result held true even when controlling for offspring feeding behavior and when analyzing data at the individual (non-aggregated) level, though the effect size decreased slightly at higher resolution due to increased within-individual noise.

C. Mediation Analysis

• Size is not the mediator: The study tested if the link was driven by body size.
  • Maternal feeding behavior did not predict maternal size at parturition.
  • Maternal size at parturition did not predict offspring size or offspring activity.
  • Offspring size at birth was linked to offspring activity, but this did not explain the maternal feeding $\to$ offspring activity link.
• Conclusion: The transmission is not mediated by body size, suggesting a more complex physiological or epigenetic mechanism.

4. Key Contributions

1. First Evidence of Non-Genetic Inheritance of Stochasticity: The study provides the first empirical evidence that phenotypic variation arising from stochastic processes (in the absence of genetic/environmental variance) can be transmitted across generations.
2. Cross-Trait Transmission Mechanism: It reveals a specific cross-trait inheritance pattern (Maternal Feeding $\to$ Offspring Activity), suggesting that stochastic behaviors may alter a mother's internal physiological state (e.g., energy reserves, hormonal profiles, epigenetic marks), which then biases offspring developmental trajectories in a different behavioral domain.
3. Challenging "Evolutionary Dead Ends": The findings suggest that clonal species, often viewed as evolutionary dead ends due to lack of genetic recombination, may possess a robust capacity for adaptation via non-genetic inheritance of stochastic variation.
4. High-Resolution Behavioral Phenotyping: The study sets a new standard for behavioral ecology by utilizing massive datasets (~19,000 hours of tracking) to distinguish subtle stochastic signals from noise.

5. Significance and Implications

• Evolutionary Theory: This work expands the scope of non-genetic inheritance. Previously, non-genetic inheritance was largely studied in the context of environmentally induced plasticity (e.g., maternal stress). This study shows that even stochastic (random) variation can be inherited, potentially acting as "non-genetic standing variation" that allows populations to adapt rapidly to environmental changes without waiting for genetic mutations.
• Mechanism of Inheritance: The lack of mediation by body size implies that the information carrier is likely state-based (e.g., metabolic regulation, endocrine profiles, or epigenetic modifications) rather than morphological.
• Future Directions: The authors propose that future research should investigate the specific molecular mechanisms (e.g., epigenetics) and test if these effects persist in more natural, resource-limited environments where activity directly impacts survival. They also suggest comparing clonal vs. sexual species to see if clonal lineages have evolved enhanced non-genetic inheritance mechanisms to compensate for genetic uniformity.

Conclusion: The study demonstrates that stochastic behavioral differences are not merely developmental noise but are heritable traits that can influence the phenotype of the next generation, offering a novel pathway for evolutionary adaptation in the absence of genetic diversity.