Seismologists use deep learning for improved earthquake forecasting

Peer-Reviewed Publication

UNIVERSITY OF CALIFORNIA - SANTA CRUZ

 

IMAGE: DAMAGE FROM A 2020 EARTHQUAKE IN PUERTO RICO. view more 

CREDIT: UNITED STATES GEOLOGICAL SURVEY

For more than 30 years, the models that researchers and government agencies use to forecast earthquake aftershocks have remained largely unchanged. While these older models work well with limited data, they struggle with the huge seismology datasets that are now available.

To address this limitation, a team of researchers at the University of California, Santa Cruz and the Technical University of Munich created a new model that uses deep learning to forecast aftershocks: the Recurrent Earthquake foreCAST (RECAST). In a paper published today in Geophysical Research Letters, the scientists show how the deep learning model is more flexible and scalable than the earthquake forecasting models currently used.

The new model outperformed the current model, known as the Epidemic Type Aftershock Sequence (ETAS) model, for earthquake catalogs of about 10,000 events and greater.

“The ETAS model approach was designed for the observations that we had in the 80s and 90s when we were trying to build reliable forecasts based on very few observations,” said Kelian Dascher-Cousineau, the lead author of the paper who recently completed his PhD at UC Santa Cruz. “It’s a very different landscape today.” Now, with more sensitive equipment and larger data storage capabilities, earthquake catalogs are much larger and more detailed

“We’ve started to have million-earthquake catalogs, and the old model simply couldn’t handle that amount of data,” said Emily Brodsky, a professor of earth and planetary sciences at UC Santa Cruz and co-author on the paper. In fact, one of the main challenges of the study was not designing the new RECAST model itself but getting the older ETAS model to work on huge data sets in order to compare the two. 

“The ETAS model is kind of brittle, and it has a lot of very subtle and finicky ways in which it can fail,” said Dascher-Cousineau. “So, we spent a lot of time making sure we weren’t messing up our benchmark compared to actual model development.”

To continue applying deep learning models to aftershock forecasting, Dascher-Cousineau says the field needs a better system for benchmarking. In order to demonstrate the capabilities of the RECAST model, the group first used an ETAS model to simulate an earthquake catalog. After working with the synthetic data, the researchers tested the RECAST model using real data from the Southern California earthquake catalog.

They found that the RECAST model — which can, essentially, learn how to learn — performed slightly better than the ETAS model at forecasting aftershocks, particularly as the amount of data increased. The computational effort and time were also significantly better for larger catalogs.

This is not the first time scientists have tried using machine learning to forecast earthquakes, but until recently, the technology was not quite ready, said Dascher-Cousineau. New advances in machine learning make the RECAST model more accurate and easily adaptable to different earthquake catalogs.

The model’s flexibility could open up new possibilities for earthquake forecasting. With the ability to adapt to large amounts of new data, models that use deep learning could potentially incorporate information from multiple regions at once to make better forecasts about poorly studied areas.

“We might be able to train on New Zealand, Japan, California and have a model that's actually quite good for forecasting somewhere where the data might not be as abundant,” said Dascher-Cousineau.

Using deep-learning models will also eventually allow researchers to expand the type of data they use to forecast seismicity.

“We’re recording ground motion all the time,” said Brodsky. “So the next level is to actually use all of that information, not worry about whether we’re calling it an earthquake or not an earthquake but to use everything."

In the meantime, the researchers hope the model sparks discussions about the possibilities of the new technology.

“It has all of this potential associated with it,” said Dascher-Cousineau. “Because it is designed that way.”

---

JOURNAL

Geophysical Research Letters

DOI

10.1029/2023GL103909 

ARTICLE TITLE

Using Deep Learning for Flexible and Scalable Earthquake Forecasting

 

New research reveals Earth's ancient ‘breath’: Study reveals connection between atmospheric changes and mantle chemistry

Peer-Reviewed Publication

UNIVERSITY OF PORTSMOUTH

 

IMAGE: GRAPH SHOWING CHANGES IN ATMOSPHERIC CONDITIONS view more 

CREDIT: DR HUGO MOREIRA

An international team of scientists have uncovered an important link between Earth’s early atmosphere and the chemistry of its deep mantle.

The study, which was led by researchers at the University of Portsmouth and University of Montpellier, sheds new light on the evolution of life on our planet and the rise of atmospheric oxygen.

The team investigated magmas formed in ancient subduction zones, where portions of Earth’s crust sink back into the mantle, from a pivotal moment in Earth's history – the Great Oxidation Event (GOE). This event, which is estimated to have happened between 2.1 and 2.4 billion years ago, was a period of time when oxygen levels in Earth's atmosphere increased rapidly and transformed life and environments on Earth.

However, there has been little research into how atmospheric changes have left their mark on the Earth’s mantle.

The new study, published in the journal Nature Geoscience, examined the role of plate tectonics – the process by which our planet's outer shell moves and reshapes its surface – in cycling and exchanging elements between the atmosphere, Earth's surface, and the deep mantle. Until now, reliable methods to understand these interactions were elusive.

By studying magmas from before and after the GOE, the team found a shift from reduced to more oxidised magmas. This was a result of the deep subduction of oxidised sediments from mountains transformed into sediments during weathering and erosion that were then recycled into the mantle via subduction processes – revealing how sediment recycling provided atmospheric access to the mantle.

This discovery implies that these ‘whiffs’ of oxygen may have changed the mantle by contributing to increased oxidation of calc-alkaline magma, altering the composition of the continental crust, and leading to the formation of ore deposits on Earth.

Lead author, Dr Hugo Moreira from the University of Montpellier and visiting researcher at the University of Portsmouth, said: “With these findings, our understanding of Earth's ancient ‘breath’ has taken a significant leap forward. Not only does it provide crucial insights into Earth's geological evolution, but it also sheds light on how the deep Earth and its mantle are intimately connected to atmospheric changes. It provides us a better understanding of the relationship between Earth's external and internal reservoirs.

“Moreover, it raises fascinating questions about the role that oxygen played in shaping our planet's history and the conditions that paved the way for life as we know it.”

The research team used the ID21 beamline at the European Synchrotron Radiation Facility in France to analyse sulphur state in minerals found in two-billion-year-old zircon crystals from the Mineiro Belt in Brazil, which acted as time capsules, preserving their original composition. They discovered that minerals from magmas that crystallised before the GOE had a reduced sulphur state. However, after the GOE, these became more oxidised.

Dr Moreira said: “Mantle oxygen fugacity, in simple terms, is a measure of oxygen's ability to drive chemical reactions in magmas and is critical for understanding volcanic activity and ore formation. However, in the past, we lacked a reliable way to track changes in this parameter for ancient parts of Earth’s history – until now.

“It provides a powerful tool for understanding the relationship between Earth's external and internal reservoirs. Sulphur speciation and magma fugacity are dynamic parameters that can change throughout a magma's journey from formation to crystallization. While our study considered factors like pressure and temperature, further analyses are needed to trace the complete ‘fugacity path’ from magma generation to final crystallisation.”

Co-author Professor Craig Storey, Professor of Geology at the University of Portsmouth, said: “Our study opens exciting new avenues of research, offering a deeper understanding of the Earth's ancient past and its profound connection to the development of our atmosphere. It challenges us to ponder questions about the evolution of magma types over time and the intricate interplay between plate tectonics and atmospheric cycles.”

Dr Moreira added: “As we continue to probe the mysteries of Earth’s geological history, one thing is certain - there is much more to discover beneath the surface.”

The study involved researchers from the University of Portsmouth, the Universities of Brest, Montpellier and University of Sorbonne, (France), the Federal University of Ouro Preto and University of São Paulo (Brazil) and the European Synchrotron Radiation Facility.

Reduced apatite inclusions

Oxidised apatite inclusions

CREDIT

Dr Hugo Moreira

CAPTION

The European Synchrotron Radiation Facility

---

JOURNAL

Nature Geoscience

METHOD OF RESEARCH

Imaging analysis

SUBJECT OF RESEARCH

Not applicable

ARTICLE TITLE

Sub-arc mantle fugacity shifted by sediment recycling across the Great Oxidation Event’

TELL THE GOP

An ‘introspective’ AI finds diversity improves performance

Peer-Reviewed Publication

NORTH CAROLINA STATE UNIVERSITY

An artificial intelligence with the ability to look inward and fine tune its own neural network performs better when it chooses diversity over lack of diversity, a new study finds. The resulting diverse neural networks were particularly effective at solving complex tasks.

“We created a test system with a non-human intelligence, an artificial intelligence (AI), to see if the AI would choose diversity over the lack of diversity and if its choice would improve the performance of the AI,” says William Ditto, professor of physics at North Carolina State University, director of NC State’s Nonlinear Artificial Intelligence Laboratory (NAIL) and co-corresponding author of the work. “The key was giving the AI the ability to look inward and learn how it learns.”

Neural networks are an advanced type of AI loosely based on the way that our brains work. Our natural neurons exchange electrical impulses according to the strengths of their connections. Artificial neural networks create similarly strong connections by adjusting numerical weights and biases during training sessions. For example, a neural network can be trained to identify photos of dogs by sifting through a large number of photos, making a guess about whether the photo is of a dog, seeing how far off it is and then adjusting its weights and biases until they are closer to reality.

Conventional AI uses neural networks to solve problems, but these networks are typically composed of large numbers of identical artificial neurons. The number and strength of connections between those identical neurons may change as it learns, but once the network is optimized, those static neurons are the network.

Ditto’s team, on the other hand, gave its AI the ability to choose the number, shape and connection strength between neurons in its neural network, creating sub-networks of different neuron types and connection strengths within the network as it learns. 

“Our real brains have more than one type of neuron,” Ditto says. “So we gave our AI the ability to look inward and decide whether it needed to modify the composition of its neural network. Essentially, we gave it the control knob for its own brain. So it can solve the problem, look at the result, and change the type and mixture of artificial neurons until it finds the most advantageous one. It’s meta-learning for AI. 

“Our AI could also decide between diverse or homogenous neurons,” Ditto says. “And we found that in every instance the AI chose diversity as a way to strengthen its performance.”

The team tested the AI’s accuracy by asking it to perform a standard numerical classifying exercise, and saw that its accuracy increased as the number of neurons and neuronal diversity increased. A standard, homogenous AI could identify the numbers with 57% accuracy, while the meta-learning, diverse AI was able to reach 70% accuracy.
 
According to Ditto, the diversity-based AI is up to 10 times more accurate than conventional AI in solving more complicated problems, such as predicting a pendulum’s swing or the motion of galaxies.

“We have shown that if you give an AI the ability to look inward and learn how it learns it will change its internal structure – the structure of its artificial neurons – to embrace diversity and improve its ability to learn and solve problems efficiently and more accurately,” Ditto says. “Indeed, we also observed that as the problems become more complex and chaotic the performance improves even more dramatically over an AI that does not embrace diversity.”

The research appears in Scientific Reports, and was supported by the Office of Naval Research (under grant N00014-16-1-3066) and by United Therapeutics. John Lindner, emeritus professor of physics at the College of Wooster and visiting professor at NAIL, is co-corresponding author. Former NC State graduate student Anshul Choudhary is first author. NC State graduate student Anil Radhakrishnan and Sudeshna Sinha, professor of physics at the Indian Institute of Science Education and Research Mohali, also contributed to the work. 

-peake-

Note to editors: An abstract follows.

“Neuronal diversity can improve machine learning for physics and beyond”

DOI: 10.1038/s41598-023-40766-6

Authors: Anshul Choudhary, Anil Radhakrishnan, John F. Lindner, William L. Ditto, North Carolina State University Nonlinear Artificial Intelligence Laboratory; Sudeshna Sinha, Indian Institute of Science Education and Research Mohali
Published: Aug. 21, 2023 in Scientific Reports

Abstract:
Diversity conveys advantages in nature, yet homogeneous neurons typically comprise the layers of artificial neural networks. Here we construct neural networks from neurons that learn their own activation functions, quickly diversify, and subsequently outperform their homogeneous counterparts on image classification and nonlinear regression tasks. Sub-networks instantiate the neurons, which meta-learn especially efficient sets of nonlinear responses. Examples include conventional neural networks classifying digits and forecasting a van der Pol oscillator and physics-informed Hamiltonian neural networks learning Hénon–Heiles stellar orbits and the swing of a video recorded pendulum clock. Such learned diversity provides examples of dynamical systems selecting diversity over uniformity and elucidates the role of diversity in natural and artificial systems.
 

---

JOURNAL

Scientific Reports

DOI

10.1038/s41598-023-40766-6 

METHOD OF RESEARCH

Computational simulation/modeling

SUBJECT OF RESEARCH

Not applicable

ARTICLE TITLE

Neuronal diversity can improve machine learning for physics and beyond

Amsterdam UMC is building models to enable greater use of AI in the health care system

Grant and Award Announcement

AMSTERDAM UNIVERSITY MEDICAL CENTERS

80% of all patient data is unstructured. Notes from a conversation with a GP, the evaluation of a specialist in a university medical centre or even a recommendation from a pharmacist. While this 'unstructured’ data is no problem for the human eye, it presents an unsurmountable challenge to an AI-algorithm. One that is "preventing AI from reaching its full potential," in the view of Amsterdam UMC, Assistant Professor Iacer Calixto. To give AI the helping hand that it needs, Calixto is set to lead a project that will "tackle the important challenges that hinder its use in clinical practice,” thanks to funding from the NWO. 

"We need to devise methods that are human-centred and responsible by design if we want these methods to be implemented in practice," says Calixto. The project will build on Natural Language Processing (NLP) techniques that already underpin the increasingly popular, ChatGPT. Currently, the unstructured nature of this data means that software such as ChatGPT cannot be easily used in the health care sector. However, the software itself offers plenty of opportunities for the sector. With promises to improve data entry, decision making and to free up crucial time that doctors and nurses can instead spend on patient care.  

Ensuring Privacy is Maintained  

"Protecting the privacy of our patients is a top priority at Amsterdam UMC, and that isn't different when we are developing, testing or using AI-algorithms," says Mat Daemen, vice-dean of Research at Amsterdam UMC. 

To ensure that AI can also be used in a safe way, this project will also address issues relating to privacy. By developing new 'synthetic' patient records, based around simulated information. These records mimic real patient records, in order to facilitate healthcare and research, while protecting the information of the 'real' patients.  

"One of the main bottlenecks of doing research in healthcare is access to high-quality data to train and validate machine learning models. Part of our project will generate synthetic patient records that include not only structured but also unstructured data such as free-text highlights from a consultation with a GP. These synthetic records, though not from real patients, can still be very useful to enable easier access to high-quality healthcare data for researchers and clinicians," says Calixto. 

Responsibly Dutch 

Another sticking point for the use of AI in the Dutch health sector, is a rather more self-evident one: language. Software such as ChatGPT are built on language databases, and these are predominantly in English. By building new models that are trained on Dutch medical records, the project will increase the reliability of existing tools as well as making them easier to use for professionals on the wards or in the treatment room.  

This is a bold project that will ensure the Amsterdam UMC is one of the forces driving innovation in healthcare with artificial intelligence and natural language processing. Results obtained in this project, for instance, synthetic patient records will benefit the entire Dutch healthcare ecosystem, including other hospitals and university medical centres, says Calixto. 

The responsibility of this AI project is not only limited to the important goal of maintaining patient privacy. The project will also seek to remove any aspects of discrimination and unfairness that may exist in existing AI models. For Daemen, this is an essential condition for the use of AI in Amsterdam UMC, and something that this project has at its core. "This project is an important addition to the efforts of many experts in Amsterdam UMC and in the Amsterdam region to introduce and use AI tools in a human centred and responsible way," he concludes.   

 

Genomic model suggests population decline in human ancestors

Peer-Reviewed Publication

AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE (AAAS)

Between 800,000 and 900,000 years ago, the population of human ancestors crashed, according to a new genomic model by Wangjie Hu and colleagues. They suggest that there were only about 1280 breeding individuals during this transition between the early and middle Pleistocene, and that the population bottleneck lasted for about 117,000 years. The researchers say about 98.7% of the ancestral population was lost at the beginning of the bottleneck. This decline coincided with climate changes that turned glaciations into long-term events, a decrease in marine surface temperatures, and a possible long period of drought in Africa and Eurasia. Hu et al. developed a coalescence model that looks at divergence between gene lineages and can be used to estimate past population size, using it to analyze genomic sequences from 3154 people from 10 African and 40 non-African populations. The ancient “bottleneck was directly found in all 10 African populations, but only a weak signal of the existence of such was detected in all 40 non-African populations,” Hu et al. write. The proposed bottleneck also coincided with the time that many researchers think the last common ancestor of Denisovans, Neanderthals and modern Homo sapiens lived, but the bottleneck theory needs to be tested against the archaeological and fossil human evidence, Nick Ashton and Chris Stringer write in a related Perspective. “If, as seems likely, humans were widespread inside and outside of Africa in the period between about 800-900,000 years BP … whatever caused the inferred bottleneck was limited in its effects on the wider non-sapiens lineage populations, or any effects were short-lived,” the Perspective authors add.

---

JOURNAL

Science

DOI

10.1126/science.abq7487 

ARTICLE TITLE

Genomic inference of a severe human bottleneck during the Early to Middle Pleistocene transition

Early ancestral bottleneck could’ve spelled the end for modern humans

Peer-Reviewed Publication

CHINESE ACADEMY OF SCIENCES HEADQUARTERS

 

IMAGE: THE CORE FORMULA OF OUR NEW INFERENCE METHOD IS SHOWN. THE IMAGE DEPICTS A CLIFF PAINTING, ILLUSTRATING THE POPULATION OF HUMAN ANCESTOR PULL TOGETHER TO SURVIVE THE UNKNOWN DANGER IN THE DARKNESS DURING THE ANCIENT SEVERE BOTTLENECK. view more 

CREDIT: IMAGE BY SHANGHAI INSTITUTE OF NUTRITION AND HEALTH, CAS

How a new method of inferring ancient population size revealed a severe bottleneck in the human population which almost wiped out the chance for humanity as we know it today.

An unexplained gap in the African/Eurasian fossil record may now be explained thanks to a team of researchers from China, Italy and the United States. Using a novel method called FitCoal (fast infinitesimal time coalescent process), the researchers were able to accurately determine demographic inferences by using modern-day human genomic sequences from 3,154 individuals. These findings indicate that early human ancestors went through a prolonged, severe bottleneck in which approximately 1,280 breeding individuals were able to sustain a population for about 117,000 years. While this research has illuminated some aspects of early to middle Pleistocene ancestors, there are many more questions to be answered since uncovering this information.

A large amount of genomic sequences were analyzed in this study. However, “the fact that FitCoal can detect the ancient severe bottleneck with even a few sequences represents a breakthrough,” says senior author Yun-Xin Fu, a theoretical population geneticist at University of Texas Health Science Center at Houston.

Researchers will publish their findings online in Science on August 31, 2023 (America Eastern Standard Time). The results determined using FitCoal to calculate the likelihood for present-day genome sequences found that early human ancestors experienced extreme loss of life and therefore, loss of genetic diversity.

“The gap in the African and Eurasian fossil records can be explained by this bottleneck in the Early Stone Age as chronologically. It coincides with this proposed time period of significant loss of fossil evidence,” says senior author Giorgio Manzi, an anthropologist at Sapienza University of Rome. Reasons suggested for this downturn in human ancestral population are mostly climatic: glaciation events around this time lead to changes in temperatures, severe droughts, and loss of other species, potentially used as food sources for ancestral humans.

An estimated 65.85% of current genetic diversity may have been lost due to this bottleneck in the early to middle Pleistocene era, and the prolonged period of minimal numbers of breeding individuals threatened humanity as we know it today. However, this bottleneck seems to have contributed to a speciation event where two ancestral chromosomes may have converged to form what is currently known as chromosome 2 in modern humans. With this information, the last common ancestor has potentially been uncovered for the Denisovans, Neanderthals, and modern humans (Homo sapiens).

We all know that once a question is answered, more questions arise.

“The novel finding opens a new field in human evolution because it evokes many questions, such as the places where these individuals lived, how they overcame the catastrophic climate changes, and whether natural selection during the bottleneck has accelerated the evolution of human brain,” says senior author Yi-Hsuan Pan, an evolutionary and functional genomics at East China Normal University (ECNU).

Now that there is reason to believe an ancestral struggle occurred between 930,000 and 813,000 years ago, researchers can continue digging to find answers to these questions and reveal how such a small population persisted in assumably tricky and dangerous conditions. The control of fire, as well as the climate shifting to be more hospitable for human life, could have contributed to a later rapid population increase around 813,000 years ago.

“These findings are just the start. Future goals with this knowledge aim to paint a more complete picture of human evolution during this Early to Middle Pleistocene transition period, which will in turn continue to unravel the mystery that is early human ancestry and evolution,” says senior author LI Haipeng, a theoretical population geneticist and computational biologist at Shanghai Institute of Nutrition and Health, Chinese Academy of Sciences (SINH-CAS).

This research was jointly led by LI Haipeng at SINH-CAS and Yi-Hsuan Pan at ECNU. Their collaborators, Fabio Di Vincenzo at the University of Florence, Giogio Manzi at Sapienza University of Rome, and Yun-Xin Fu at the University of Texas Health Science Center at Houston, have made important contribution to the findings. The research was first-authored by HU Wangjie and HAO Ziqian who used to be students/interns at SINH-CAS and ECNU. They are currently affiliated with Icahn School of Medicine at Mount Sinai, and Shandong First Medical University & Shandong Academy of Medical Sciences, respectively. DU Pengyuan at SINH-CAS, and CUI Jialong at ECNU also contributed to this research.

The African hominin fossil gap and the estimated time period of chromosome fusion is shown on the right.

CREDIT

Image by Science

JOURNAL

Science

DOI

10.1126/science.abq7487 

ARTICLE TITLE

Genomic inference of a severe human bottleneck during the Early to Middle Pleistocene transition

 

Mutation rates in whales are much higher than previously reported

Estimate lowers the pre-exploitation number of North Atlantic humpback whales by 86 percent

Peer-Reviewed Publication

UNIVERSITY OF GRONINGEN

 

IMAGE: THIS PICTURE SHOWS HOW A SMALL BIOPSY IS COLLECTED, USING A HOLLOW-POINT CROSSBOW DART, FROM A FIN WHALE IN THE GULF OF MAINE. THE FIN WHALE IS ONE OF THE SPECIES INCLUDED IN THE STUDY. view more 

CREDIT: CENTER FOR COASTAL STUDIES IMAGE COLLECTED UNDER U.S. NMFS ESA/MMPA PERMIT 1632 (PERMIT NUMBER MUST BE INCLUDED IN CREDIT!)

An international team of marine scientists, led by the University of Groningen in the Netherlands and the Center for Coastal Studies in the USA, has studied the DNA of family groups from four different whale species to estimate their mutation rates. The results revealed much higher mutation rates than previously thought, and which are similar to those of smaller mammals such as humans, apes, and dolphins. Using the newly determined rates, the group found that the number of humpback whales in the North Atlantic before whaling was 86 percent lower than earlier studies suggested. The study is the first proof that this method can be used to estimate mutation rates in wild populations and was published in the journal Science on 1 September.

Mutation rate is a key parameter in genetics and genomics, where it is used to determine rates of evolution and adaptation. It is also used to derive the number of whales in the oceans before they were decimated by large-scale commercial whaling. However, estimating the rate at which new mutations appear in whales, or in any wild species, is difficult.

Pedigree method

For a long time, the phylogenetic method was used to measure mutation rates. This method uses fossil data from different species to estimate when they diverged. Subsequently, DNA from those species is compared to infer how many mutations must have occurred since the divergence. ‘However, the fossil record is not that exact. And some mutations may have disappeared over time,’ says Per Palsbøll, Professor of Marine Evolution and Conservation at the University of Groningen. He has studied whales since the late 1980s and is a corresponding author of the Science paper.

A more recent approach is the pedigree method, which uses the genomes of a pair of parents and their offspring to identify new mutations in the offspring. This more direct method relies on very few assumptions and is ideal for comparing mutation rates among different species, such as whales and humans.

Especially in wild species, the challenge is to obtain tissue samples from both parents and their offspring. First author Marcos Suárez-Menéndez: ‘The method has only been used on a handful of animals that are living in the wild, such as a single wolf pair and their cubs. It has also been used to estimate mutation rates in zoo animals, although it is uncertain if this reflects the mutation rates in the wild where the conditions are very different.’ However, the team, comprising scientists from the Netherlands, USA, Greenland, Denmark, Canada, and the UK, were able to use skin biopsy samples collected from whales during a collaboration that has been ongoing for more than thirty years.

Crossbow

Palsbøll collected his first whale biopsy samples amongst icebergs in the waters off West Greenland in 1988. ‘To do this, we had to sail very close to a whale and then fire a dart with a hollow point using a crossbow.’ The dart punches out a sample and bounces back into the water from where it is collected.

Finding both parents of a whale calf is the first step in measuring the mutation rate using the pedigree method. This is where large-scale DNA analyses come in. Suárez-Menéndez analyzed data that were generated by the other first author, Martine Bérubé, from microsatellite markers in DNA. This DNA was extracted from a large archive of whale biopsy samples and used to create a genetic fingerprint of individuals. ‘I sifted through the microsatellite data to find individuals that were related as mother and calf. Next, I looked for possible fathers in the database.’

In this way, he managed to identify 212 putative parent and offspring trios in four different whale species. The DNA of eight trios was then sent off for genome sequencing. After a final paternity check, Suárez-Menéndez and his colleagues estimated the number of new mutations in the calf and the average mutation rate in whales.

Industrial whaling

The results showed that the mutation rates in whales are similar to the rates seen in pedigrees in smaller mammals such as humans, apes, and dolphins. In contrast, earlier estimates in whales using the phylogenetic method were much lower compared to these smaller mammals. Suárez-Menéndez: ‘And just like in humans, most new mutations originate from the father. So, whales are very similar to us in this respect.’

The team also used a slightly different maternal pedigree method to estimate the mutation rates in DNA from mitochondria, the cell’s power plants. This method has so far only been used in penguins. Mitochondria and their DNA are passed on through the maternal line and Suárez-Menéndez took advantage of four decades of sighting data of humpback cow and calf pairs in the Gulf of Maine, directed by senior author Jooke Robbins at the Center for Coastal Studies. ‘Our study revealed that the mutation rate in whale mitochondrial DNA is also much higher than earlier estimates based on the phylogenetic method,’ explains Suárez-Menéndez.

The newly determined, higher mutation rates were used to infer that the number of whales in the North Atlantic before industrial whaling. The result was 86 percent lower than earlier reported estimates based on phylogenetic mutation rates. ‘Our new mutation rates suggested that some 20,000 humpback whales lived in the North Atlantic before commercial whaling, in contrast to the previous estimate of 150,000,’ says Palsbøll. This is important information, not only for the conservation of whales but also for our understanding of the state of the oceans before whaling. Palsbøll: ‘Another conclusion of wide-ranging consequences is that our study shows that it is entirely feasible to estimate the mutation rate in wild animals.’

Cancer

The human-like mutation rates in whales also led the authors to reject one possible cause of Peto's paradox. This is the observation that, at the species level, the incidence of cancer does not appear to correlate with the number of cells in an organism. Whales have a hundred to a thousand times more cells than, for example, humans, so if they have the same cancer rate as humans, they should get cancer very early in life. Several mechanisms have been proposed for protecting these large sea mammals against cancer. One of those is a slower mutation rate as a consequence of whales having much lower metabolic rates. The discovery that this is not the case, implies that other mechanisms are probably at play in whales, such as an increase in the number of copies of the p53 gene which protects against cancer.

Finally, as the study relied on a large number of tissue samples that have been collected over several decades, the Science paper highlights the importance of long-term ecological research projects. Palsbøll: ‘It is difficult to acquire sustained funding for these kinds of long-term ecological studies. However, we wouldn’t have been able to do this research without the sustained commitment and dedication of the many colleagues who recorded all the sightings and collected the samples that our study relied on.’

Reference: Marcos Suárez-Menéndez, Martine Bérubé, Fabrício Furni, Vania E. Rivera-León, Mads-Peter Heide-Jørgensen, Finn Larsen, Richard Sears, Christian Ramp, Britas Klemens Eriksson, Rampal S. Etienne, Jooke Robbins, and Per J. Palsbøll: Wild pedigrees inform mutation rates and historic abundance in baleen whales. Science, 1 September 2023

Nuclear mutation rates were determined through genome sequencing of whale trios (mother, father, and calf). The mutation rate was calculated by counting nucleotide base changes (i.e. mutations) exclusive to the calf and dividing them by the total number of bases. The mitochondrial mutation rate was estimated by analyzing the transmission of mitochondrial heteroplasmy (two distinct mitochondrial genomes within an individual caused by a mutation) and its frequency across hundreds of maternal lineages.

CREDIT

Marcos Suárez-Menéndez, University of Groningen

---

JOURNAL

Science

DOI

10.1126/science.adf2160 

METHOD OF RESEARCH

Observational study

SUBJECT OF RESEARCH

Animals

ARTICLE TITLE

Wild pedigrees inform mutation rates and historic abundance in baleen whales

ARTICLE PUBLICATION DATE

1-Sep-2023