Ancient Egyptians identified three seasons based on cycles of the Nile River: inundation, emergence and harvest. In tropical countries, it’s the rain that tends to divide the year in two: a wet and a dry season. Elsewhere the calendar demarcates four seasons: autumn, winter, spring and summer.


But now, researchers at Stanford University have found that human biology, rather than rivers, rainfall or calendars, could be used to determine the seasons. In their study, published in the journal Nature Communications in October 2020, the Stanford researchers discovered our bodies seem to set their own rhythm, splitting the year into two seasonal time periods.

Or at least that’s the case if you live in California, where the study was carried out. Since every geographical location has unique environmental conditions, their approach may be used the count the seasons in other parts of the world too.

“People say there are four seasons of three months each. But why four? There could be 15 or could be 2. Why don’t we let biology tell us?” asks Prof Michael Snyder, principal investigator of the study.

To determine the human seasons, Snyder’s team profiled the biology of 105 volunteers in the San Francisco Bay area over a period of four years. They regularly sampled and measured tens of thousands of molecules and microbes from the participants’ blood, noses and guts. This type of study is called ‘deep longitudinal multiomics profiling’.

On sample days, the researchers also collected meteorological data (such as air temperature and solar radiation) and airborne pollen counts.

This massive effort was undertaken to create a better picture of how the changing seasons might be affecting our physiology and health.

Read more about the biometrics:

What is multiomics?

When added to a molecular or microbial term, the suffix ‘-omics’ refers to the comprehensive analysis of a collection of those molecules or microbes. For example, genomics is the comprehensive study of all of an organism’s genes. Genomics is different from genetics, which considers single genes or their variants.

More like this

Scientists often talk informally of ‘omics’ or ‘omics groups’, which may include genomics, metabolomics, proteomics, transcriptomics, epigenomics, microbiomics and others.

Multiomics, which formed the basis of the study led by Snyder, is a branch of molecular biology in which researchers combine and analyse large data sets representing different omics groups. The goal of multiomics is to highlight relationships among the collections of molecules and their functions.

How many seasons does the body have?

After four years of testing poo, taking blood samples and logging the weather, the team used powerful computational tools to try and find patterns between the volunteers’ biology and their environment. What they found surprised them.

There were two signals. One was a group of molecules that seemed to peak in December – a season the researchers dubbed late fall/early winter. This included markers related to immune responses such as the complement system, a collection of proteins that work together to eliminate infectious microorganisms, which peaked during this time. Unsurprisingly, this correlated with the period we know viral infections are also high.

The second signal, however, did come as a surprise.

“I thought the other [season] would be in June or July when it’s pretty hot, but that wasn’t true,” Snyder says. Instead, the second season peaked in late April – a season they called ‘late spring’.

This season’s peak made sense in hindsight, as late April also corresponded with a time of high pollen counts at the end of California’s rainy season. The pollen caused a reaction in a large enough subset of people to contribute to the seasonal peak in the immune response.

Pollen (seen here under a scanning electron microscope) peaks in late April in California, just as some of the participants’ immune systems reached a second high point © Getty Images
Pollen (seen here under a scanning electron microscope) peaks in late April in California, just as some of the participants’ immune systems reached a second high point © Getty Images

The findings added nuance to earlier understandings regarding how human biology interacts with seasonal patterns in the presence or absence of disease. For example, scientists had known that the disease risk marker HbA1c (an indicator of recent average blood glucose levels) was often higher in winter than in summer for diabetic patients.

What they didn’t know was how levels varied throughout the year for non-diabetic patients. This study revealed that participants in general – diabetic or not – experienced peak HbA1c levels in late April. Snyder made sense of this by noting that late April is a time when people emerge from a somewhat more dormant period of not exercising as much.

The team also observed that PER1, a gene responsible for circadian rhythms, had a seasonal pattern, with its highest expression in spring. Furthermore, other studies have found that PER1 may play a role in the development of cancer and that incidences of localised tumours appear to be highest in the spring.

Snyder’s team suggests that their observation of a spring peak for PER1 provides additional evidence that the gene may contribute in some way to cancer growth.

An individual approach to health

So how could this research be useful? For a start, it will help us understand the fluctuations in the human body from patient to patient – variations that can’t be measured in the tests we typically get when we visit the GP.

But also, a standalone measure, such as a temperature reading, is often interpreted against a population average, without the context of the individual’s normal, healthy baseline. For example, the average human temperature is 36.5°C, though even that varies by gender, age and throughout the day, according to a Journal Of Internal Medicine study.

The San Francisco Bay area © Getty Images
The study was carried out in the San Francisco Bay area, but the methods could be used for people in other regions © Getty Images

But an individual’s ‘normal’ temperature may range from 36.1 to 37.2°C. A patient with a low baseline who registers a temperature that’s towards the high end of the normal range may indeed have a fever, even though the doctor taking that patient’s temperature may not realise it.

“People usually go to the doctor when they’re ill. They don’t often go when they’re healthy. We never take advantage of the longitudinal nature of data – collecting data over time. That’s the essence of what we’re trying to do,” Snyder says.

Besides, an individual’s temperature or other health metrics may vary over a given year, even when they’re healthy. And disease markers for conditions such as arthritis, sleep disorders, and many neurological and psychiatric illnesses may also vary throughout the year. All of which raises questions about seasonal influences on health.

“If your cholesterol is higher in the winter than in the summer, is that normal biological variation or is that signifying a potential health problem?” asks Dr Laura Cox, professor of molecular medicine at Wake Forest School of Medicine, who was not involved in the study.

Read more about the future of medicine:

Personalised health

Until nearly the end of the 20th Century, researchers did not understand the existence or usefulness of a multiomics approach to human biology. Instead, they conducted targeted studies that looked at the influence of a single gene or a single protein on health or disease.

In recent decades, however, multiomics such as genomics and proteomics (collections of genes or proteins, respectively) have allowed researchers to gain a more integrated understanding of biological impacts on health and disease.

“The genes and the proteins and the metabolites and the lipids are all talking to each other all the time,” Cox explains. By quantifying tens of thousands of these measurements in a deep longitudinal multiomics profile, researchers may then determine what is, and is not, likely to be biologically important.

Snyder’s study offers personalised health models – one for each study participant – that follow and predict health trajectories. The models paint a picture of the normal biological variation in the different omics groups for the patient throughout the year, which is key in catching disease in its early stages.

Evidence supporting this approach is compelling. Snyder reports that the deep longitudinal multiomics profiles uncovered major health discoveries among nearly half of the participants, including early diagnoses of lymphoma, heart problems and
a BRCA gene mutation indicating a high risk for breast cancer.

Snyder himself is a devout health logger. He even published results from his personalised omics profile in a 2012 paper in the journal Nature Reviews Genetics (he was both an author and the study’s only participant).

Study leader Prof Michael Snyder with several wearable health devices © Steve Fisch/Stanford University
Study leader Prof Michael Snyder is a fan of tracking his health – he wears eight devices every day. He’s keen for personalised health data to be more widespread © Steve Fisch/Stanford University

Today he wears eight portable devices to track his daily health, including four smartwatches, a continuous glucose monitor, a meter that measures environmental exposures, a health tracker ring, and a pulse oximeter. Recently, his smartwatch and pulse oximeter indicated that his blood oxygen level had dropped at the same time that his heart rate had increased, which turned out to be the first indication of a forthcoming diagnosis of Lyme disease.

The seasonal study that Snyder and his team have carried out is a research version of personalised health medicine that represents deep profiling, but he accepts that it will be difficult to do for the entire population.

“You’re not going to do that for everybody. But we can try and figure out what’s most useful and then try to put out a cheaper, higher utility, ‘most bang-for-the-buck’ version,” Snyder says. He also mentions that his ultimate goal is to use big data to build personalised health models depicting trajectories for every single person on the planet.

Cox, who has no connection to Snyder, calls the study a “tour de force”, as few others have studied as much multiomics data over such a long a period of time. “Often, we see a very brief snapshot in time and infer a whole, continuous timeline based on that one brief snapshot,” she says. “It brings up: how much are we missing?”

Preventative care

It’s clear to see how tracking someone’s every bodily fluid, microbe and molecule for four years would create a higher resolution picture of their health, and allow doctors to take preventative measures protect them from disease. Sadly, however, this won’t be manageable in the real world.

For a start, the culture of medicine would need to shift its focus from the current model of diagnosing and treating someone according to their symptoms, to focusing on early diagnostics and prevention. This is assuming that the high cost of comprehensively tracking everyone’s health metrics – not to mention any privacy concerns – could be addressed.

Nonetheless, in revealing that, according to their biology, northern Californians experience two seasons rather than the traditional four we associate with the calendar, this research suggests that seasonal influences should be considered when addressing human health and disease management.

Plus, the study offers a template for identifying seasonal counts and influences in other parts of the world, which may have an impact on our understanding of human health and disease management in those regions.

So how many seasons are there? Well, it’s difficult to say for certain, but your health likely depends on the answer. “Even so,” Snyder notes, “I predict that there aren’t going to be four seasons with three months each.”



Susan is a writer and mathematician.