Zurich-led team builds a global DNA atlas of microscopic life in 362 lakes

Conventional DNA surveys of water use a molecular bait that catches only the groups it was built for, so other organisms slip through ^[9]. This team dropped the bait. Lake water is filtered, everything caught on the filter is sequenced at once, and a machine-learning tool picks out the mitochondrial DNA ^[9][10]. Choosing mitochondria is the clever part of the design. Their genomes are small, so they can be reconstructed from a mixed sample. They also generally change faster over evolutionary time than nuclear DNA, so even close relatives carry clearly different versions ^[10].

Species in the atlas are defined by a cut-off: two mitochondrial genomes at least 98.1% identical count as one species ^[12]. The rule keeps identifications consistent across lakes, seasons and studies ^[12]. It also means the roughly 3,660 species, and the six- to 20-fold excess over the leading reference database, are counts of genetic clusters drawn with that line, and they would shift if the line moved ^[6][7][12].

Of those 3,660 species, about 750 matched no known group, close to one in five ^[16]. "This shows just how large the gaps in our knowledge still are," lead author Lucas Serra Moncadas said, according to phys.org ^[11].

For water managers the useful question is toxicity. "Under the microscope, a harmless alga and its toxic relative can look identical," said Adrian-Stefan Andrei, who heads the Microbial Evogenomics group at the University of Zurich's Limnological Station and was the study's principal investigator ^[3][15]. "Their DNA likely holds the key to telling them apart." The word "likely" is accurate. The report does not describe a test showing that a toxic strain and its harmless look-alike fall on opposite sides of the 98.1% line ^[12].

The Oder River shows what a missed bloom costs. In 2022 a toxic microalga bloomed there, on the German-Polish border, and killed around 360 tonnes of fish ^[13]. The researchers say DNA-based monitoring could allow faster responses to overheated or polluted water and catch cases like the Oder's early ^[18]. "For more than 60 years, we have seen only a fraction of the life in the lakes and reservoirs that provide drinking water to more than 180 million Europeans," Andrei said ^[19]. "Soon, a single bottle of lake water could reveal who lives in it, what's new and what's changing."

I think the atlas succeeds at its first task: supplying the genetic reference data that, until now, was largely missing ^[14]. Retiring the 1958 microscope count would take a second study ^[2]. It would need the same samples counted by eye, read by sequence and tested for toxins, with the agreement between methods reported.