Researchers at the Carnegie Institution for Science, Purdue University, and Johns Hopkins University have created a test that quickly assesses materials for signs of life. Though their model has only been tested with Earth-based samples so far, the hope is that it will be used to sort through specimens obtained by the Mars Curiosity rover.
The test uses a combination of pyrolysis gas chromatography and electron impact ionization mass spectrometry called Pyr-GC-El-MS. Pairing these methods allows scientists to break down small fragments (weighing roughly 100 micrograms) of a sample to observe which compounds the sample is made of, regardless of those compounds’ molecular weights. This makes the test “agnostic” or able to be used on biotic (life-derived) or abiotic (unrelated to life) samples.
Electron mass spectrometry results are traditionally fed into a database, where they’re matched with known compounds. But when the goal is to locate life on other planets, known compounds are only so useful—after all, alien biology might involve compounds we haven’t observed here on Earth. Following this reasoning, the researchers’ test doesn’t look for specific molecules; instead, it seeks out tiny differences in the distribution and diversity of molecules and their compounds. The idea is that the patterns—not the molecules that comprise them—found within terrestrial biotic samples are similar to those found on other planets.
Identifying and comparing those patterns requires extensive work, making it an ideal job for AI. The researchers used Pyr-GC-EI-MS to analyze 134 known carbon-bearing samples, from teeth, coal, shells, and oil shales to carbonaceous meteorites and pure synthetic chemicals. Then, they trained a machine learning model to map out the molecular patterns found in each. To test the model, the researchers introduced new samples. The model could discern biotic and abiotic samples with 90% accuracy and even began sorting biotic samples into “living” and “fossil” categories.
“Biological systems on other worlds might not produce identical, or even broadly similar, suites of organic molecules to those found in modern terrestrial biology,” the researchers write in their paper for Proceedings of the National Academy of Sciences (PNAS). “Rather, we suggest that even alien biochemical systems that might differ significantly from Earth’s biochemistry would still display molecular frequency distributions that are distinct from those of background abiotic synthetic processes…We suggest that these types of differences between biotic and abiotic molecular suites can be detected and quantified using our techniques.”
The model has already been adapted for spaceflight missions, hoping it might someday help rovers look for signs of extraterrestrial life.
