What's really in a fish tank?

What do you get when you sample a fish tank for DNA?
03 December 2019




In November 2019, Naked Genetics presenter Phil Sansom threw down the gauntlet before genome-sequencing giants illumina and challenged them to identify a mystery animal by its genetic fingerprint. But this was no ordinary challenge, because all the team at illumina had to go on was a sample of water from the creature's tank. The full story is available as a Naked Genetics Podcast episode. Below is what illumina found when they tackled the water in Phil's jam jar...


This water seems to be from a fresh water tank containing cichlids, perhaps zebra mbuna and tilapia. They seem to be treating or preventing an outbreak of Aeromonas hydrophila, that causes fin rot, by using ginger in the water. The other DNA in the tank suggests there are a lot of people around and perhaps some rodents, most likely mice. Pseudomonas aeruginosa and hookworms in the water mean people should wear gloves when handling the water and certainly not drink the water as these are both human pathogens. Bacteria were detected in the water that are involved in nitrogen fixation, an important part of maintaining a healthy fish tank.


The DNA from the fish tank was analysed by comparing each of the reads to a database that contains every species that has ever been sequenced completely. The analysis is very fast and it took less than an hour to check each of the 500 millions reads against the database. However, as the database is very large it takes a large computer with over 150Gb of RAM and 50 CPUs to run the analysis. This is about the same as 8 MacBook Pros.

Most of the DNA in the tank is unknown - in fact 98% - and has not been investigated further. There are some databases, that contain more species and results from environmental sequencing, which might have been able to be classified more. Also, with relaxed parameters there is up to about 20% classified DNA; but presumably with a lot more false positives. There are low levels of golden eagle with the relaxed parameters, for example, which pretty definitely is not in there! However, the top results are unlikely to change, there are still a lot of reads to work with.

Of the DNA that can be identified about 95% is bacteria, with about 5% animal and 1% plant, and a tiny fraction of fungi and archaea.


The most common animal species in the tank is human. It is unclear why the human content is so high. Even so, there is only enough DNA for a 0.04x coverage of the human genome. The next most common are fish species of the family Cichlidae called cichlids. This would be the best guess for the contents of the tank, especially as cichlids are a common aquarium fish.

Cichlids are a wide family and contains several sub-genuses. The following are the most likely genus and species to be found in the tank. The genus is probably correct, but the species could well not be, as it depend which species have been sequenced or not. The list goes on further - this is just the top few.


A Degeni cichlid among the rocks (Haplochromis degeni)

Most of the Cichlids found originate from the African great lakes, particularly Lake Malawi.

There is low level DNA from a number of other animals, but at far lower levels than the cichlids. These include Oryzias (ricefishes), Cyprinus (Carp) and even Pygocentrus (Piranha)! There is also some mouse or rodent DNA; perhaps there is a pet store at the garden centre too? Slightly more gross is some nematode species that are parasitic hookworms! Don’t drink the water!


The main plant DNA found in the tank is ginger (Zingiber officinale). This seemed a little strange, but it is given to Nile tilapia (Oreochromis niloticus) to prevent infection by Aeromonas hydrophila, which causes them to develop ulcers, tail rot, fin rot, and hemorrhagic septicaemia. We also see that Aeromonas is one of the most common bacterial genus seen in the tank, as well as seeing Nile tilapia DNA in the tank. So this actually makes a lot of sense.


The most common bacteria is Ralstonia, which has been previously seen to be a contaminant of DNA extraction kits and is probably not from the tank.

Most of the bacteria are commonly found in fresh water environments, there are also some pathogenic species, which presumably explain the ginger. Then there are some nitrogen fixing bacteria, which are important for the nitrogen cycle in the water.

SphingopyxisSphingopyxis alaskensisIsolated from Alaskan Water
NovosphingobiumNovosphingobium pentaromativoransIsolated from estuarine sediment
SphingobiumSphingobium yanoikuyaeCommonly isolated from soil
PlesiomonasPlesiomonas shigelloidesIsolated from freshwater and freshwater fish
PorphyrobacterPorphyrobacter neustonensisPorphyrobacter occur in freshwater environments.
PseudomonasPseudomonas aeruginosaIt is found in soil, water, skin flora, and most man-made environments throughout the world.

Isolated from aquarium plants.

Don’t drink the water and wear gloves!

AeromonasAeromonas hydrophila

Aeromonas salmonicida

Pathogenic bacteria to many fish species
MesorhizobiumMesorhizobium opportunistumNitrogen fixing bacteria
BradyrhizobiumBradyrhizobium oligotrophicumNitrogen fixing bacteria

And many, many more!

So... did they get it right?

The correct answer for what was in the fishtank was a piranha. Specifically, four young red bellied piranhas - too young to be too dangerous.

However, the aquarium circulated water within individual systems, each system containing many different tanks. In this case, the piranhas were in a tropical system - along with a large number of cichlids. So it seems that illumina were exactly right, and even managed to identify the smaller traces of piranha in the system.

It is possible that the rodent DNA came from the piranha food. The food is often sold as generic pellets, so may well contain rodent. And piranhas are notoriously messy eaters, which explains why bits of food were left in the water.

The ginger in the water remains a mystery. It might be in a treatment used in the water to prevent hair algae growing. 


Add a comment