Russia’s ongoing invasion of Ukraine dominates the news and deservedly so. It will also have major repercussions on the scientific world as cooperation with Russia in many fields has stopped, leading to a lot of research work being lost including projects in Ukraine and on the International Space Station. But scientific work in other countries goes on.

• The scariest and most widely shared bit of news this month is surely the announcement of how easily AI technologies used to generate promising molecule candidates can be turned towards making biochemical weapons. Normally the machine learning software searches for molecule candidates that score well on whatever bioactivity they are targeting while minimizing toxicity. It was a simple matter then to invert the parameters to search for molecules with high predicted toxicity and very quickly the model came up with the already known nerve agent VX as well new molecules that aren’t known to exist yet but are predicted to be even more toxic. Of course all this is in simulation so it may not even be possible to synthesize these molecules and the researchers aren’t releasing their data to the public, but this was done using open-source software and publicly available datasets of toxicity and so should be easy to replicate.
• One of my favorite science stories this month is the invention of a device to detect Parkinson’s disease early. What is fascinating is how this came to be. A retired nurse in Scotland claimed to be able to smell a distinctive odor emitted by people with the disease, beginning when her own husband developed it. This ability was confirmed when she met more patients with the disease in the support groups her husband went to. A team of researchers then built a device to detect the same organic compounds that this woman could smell in the natural oils produced by the skin of patients and found that it worked. However the woman’s nose still has significantly highly accuracy.
• Next we cast our eyes to far larger objects and events in this article discussing how two supermassive black holes will collide about 10,000 years from now. They are some 9 billion years away from us of course so even if the collision is expected to cause warps in the fabric of space and time, we only detect them as ripples and the real story here is the process by which astronomers figured out what is going on.
• Finally here’s an older paper published in a journal about religious studies which I usually avoid but this one is interesting enough to be worth noting. The subject is Turkey and the decades long effort to Islamize the country. They found that such metrics as mosque attendance, trust in clergy and so forth have actually declined, probably because by associating the state so closely with the religion, the failures of the state also become the failures of religion. I believe that this is an important finding in examining repressive states which attempt to use religion to legitimize their rule.