Scientists develop new artificial intelligence method to look for life on other planets

They presented their work to scientists at the Goldschmidt Geochemistry Conference in Lyon in July this year. After receiving a positive reception from other scientists at the time, details of the system were published in the peer-reviewed journal PNAS on Monday.

“This is a significant advance in our abilities to recognise biochemical signs of life on other worlds. It opens the way to using smart sensors on unmanned spaceships to search for signs of life,” said lead researcher Robert Hazen of the Carnegie Institution’s Geophysical Laboratory and George Mason University, in a press statement.

Scientists have known since the Miller-Urey experiment in the 1950s that mixing simple chemicals in the right conditions can form some of the more complex molecules that are required for, like amino acids. Since then, many of the more complex components needed for life as we know it, like nucleotides needed to make DNA, have been detected in space.

But, how do we know whether these molecules are of biological origin or if they were made by other unknown processes over time? Knowing that is essential to deciding whether we have detected life or not.

The “life test”

The scientists used the pyrolysis gas-chromatography mass-spectrometry methods (GCMS) to analyse 135 different carbon-rich samples. These samples were taken from living cells, age-degraded samples, geologically processed fossil fuels, carbon-rich meteorites, and laboratory-synthesised organic compounds and mixtures.

As many as 59 of these samples had a biological origin, like a grain of rice, a human hair, crude oil, etc. And 74 of them were of non-biological origin like lab-synthesised compounds or samples from carbon-rich meteorites.

The samples were first heated in an environment without oxygen—a process called pyrolysis. This caused the samples to break down. After that, they were analysed in a GCMS, a device that separates the mixture into its component parts and identifies them.

They then used some machine-learning methods to train models on the three-dimensional data (time/intensity/mass) from each sample as training or testing subjects. After testing this model on samples, it was found to have an accuracy greater than 90 per cent.

The implications

“There are some interesting and deep implications which flow from this work. First, we can apply these methods to ancient samples from Earth and Mars, to find out if they were once alive. This is obviously important for looking at whether there was life on Mars, but it can also help us analyse very ancient samples from Earth, to help us understand when life first began,” explained Hazen.

According to Hazen, this also means that at a deep level, biochemistry and non-biological chemistry are somewhat different. This also suggests that if we find life elsewhere, we can tell if life on Earth and other planets came from a common origin, aligning with the theory of panspermia, or whether they have fundamentally different origins.