7 ms·
Scientists discover new antibiotics using AI
- RcouF1uZ4gsC 3y agoTigecycline https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1426172/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1426172/ is a new class of antibiotics discovered around 2005 that also works against resistant bacteria such as MRSA and VRE. So while it is great they are coming up with new antibiotics, it is by no means the first in 60 years.
- pajko 3y agoCeftaroline https://pubmed.ncbi.nlm.nih.gov/20334458/ https://pubmed.ncbi.nlm.nih.gov/20334458/ Discovered about at the same time or later. Also works against MRSA.
- PlasmonOwl 3y agoHad the same issue on Reddit recently. Toxibaccin was discovered in soils in 2018. Hell I know someone who developed one recently at a university. It’s AMAZING AI can do this, but it certainly doesn’t destroy prior approaches. Yet. Disclaimer: spelling and name of antibiotic may be incorrect. It is not my field, and the spelling is phonetic from myself. If pressed, I will ask the scientist for the specific name.
- maxerickson 3y agoI expect you are referencing https://en.wikipedia.org/wiki/Teixobactin https://en.wikipedia.org/wiki/Teixobactin
- 29athrowaway 3y agoDiscovered in some random dirt from Maine.
- hackernewds 3y ago"Discovered" like penicillin? It irks me when it's presented as accidental, when it does go through testing after
- 29athrowaway 3y agoPenicillin was also accidental? I don't understand the point.
- PlasmonOwl 3y agoMaybe he conflates discovery with accident. Materials science searches chemical space and makes discoveries. There is no accident in that kind of discovery?
- maxerickson 3y agoSeems that is a tetracycline derivative? A class of antibiotics generally refers to the mechanism of action, so I'd think it wouldn't be a new class if it is derived from an existing one.
- awwaiid 3y agoMaybe this is a missing-comma type issue, and they mean that the last time that AI came up with an antibiotic was 60 years ago
- augustulus 3y agowhere would you put the comma to fix that?
- lacrimacida 3y agoIt’s the first comma in click-bait.
- hackernewds 3y agonow for them to make one against viruses
- dvh 3y agoI've been under the impression that we don't have new antibiotics because it is not economically viable. When you invent new weight loss drug you can sell millions of pills to anyone. If you invent new antibiotic it is immediately classified as "reserve antibiotic" and the sale is restricted.
- ac2u 3y ago>it is immediately classified as "reserve antibiotic" and the sale is restricted. Is that so medical professionals can break it out as a means of last resort to save a patient and ensure the course is finished? If so that mightn't be a bad idea before we destroy the potential of new antibiotics too by having the general public misuse them render them useless over time.
- zbrozek 3y agoIt's probably a good idea. But doing that alone--without substituting market incentives with something else--means that there won't be new products to use in that way.
- garrisonhh 3y agoI think the above commenter is just pointing out how the fundamental conflict between these two things results in significantly lessened incentive for pharmaceutical companies to put money into researching new antibiotics
- schneems 3y agoI extremely know nothing about this field. Would a naive approach would to have a coordinated “crop rotation” type tactic where all hospitals switched to a primary antibiotic every C years/months do anything?
- RiDiracTid 3y agoEmpirically, many bacteria seem to be able to acquire resistance in a way that doesn't significantly impact their fitness, meaning they can basically get new resistances and keep old ones for almost arbitrarily long, so the crop rotation idea would fail massively. https://academic.oup.com/femsre/article/35/5/901/2680377 https://academic.oup.com/femsre/article/35/5/901/2680377
- photochemsyn 3y agoThe nature article is paywalled so it's impossible to see what they actually did, and the abstract is not very clear either. What benefit did they get from whole-genome sequencing, for example? Additionally, it can't be called an antibiotic until it goes through the necessary human clinical trials, where all kinds of problems can crop up - they don't seem to provide any evidence that their AI-enabled scan for human cell toxicity actually worked. Finally, the real problem with antibiotic resistance is that microbes always evolve resistance over time if the antibiotics are overused (i.e prophylactic use in industrial animal agribusiness, and careless prescriptions of antibiotics from doctors).
- throwawaymaths 3y agoAn antibiotic is an antibiotic whether or not it would kill a patient too.
- ycui1986 3y agobleach is also antibiotic then
- hackernewds 3y agoso is lava then
- kbelder 3y agoFair, but finding new antibiotics then becomes trivial.
- throwawaymaths 3y agoIsolating a new antibiotic (even if it wouldn't work on humans) is typically worth a chemistry PhD. Call me when you get it.
- mminer237 3y ago
- rambambram 3y agoWhat is AI? I've only heard of LLM.
- catchnear4321 3y agoit is just the commercial name for llm and related technologies.
- dartos 3y agoThat’s just not true. The technology that’s behind LLMs have been in use in many disciplines for decades. (Under the names “machine learning,” “statistics,” and now AI) LLMs are just the ones that are popular right now because even non technical people can interact with them. In the mid 2010s it was image classification, for example.
- skellington 3y agoSort of. The deep-learning AI revolution started with back propagation techniques (80s) that made neural nets trainable. But since then it's taken a long time to get to the relatively recent architectural breakthroughs like transformers (2017) and modern-diffusion (2020) which lead to chat-gpt and dall-e. All of the big AI breakthroughs now are based on neural nets vs classical machine learning which grouped a bunch of statistical, regression, and optimization type models together. Modern AI is the synthesis of new techniques and old ideas plus huge datasets and huge computing power.
- catchnear4321 3y ago> LLMs are just the ones that are popular right now because even non technical people can interact with them. this was essentially my point, that a /s might have made more clear, and made cost less internet points. ai and how the term is used commercially are very different things. the latter tending to be a myopic view of the former that is constrained by dollar-seeking. ai is getting to be old as shit. ai hype is in its latest wave of commercialization. llms being the current so hot right now.
- 3y ago
- pfisherman 3y agoThe headline is sensationalized. They did a virtual screen of a large compound library. QSAR / QSPR / Virtual Screening has been around since the late 1950s. The secret sauce here was the large experimental dataset - on the order of 10^5 compounds - they generated to train the model. Maybe the explainability part is kind of novel for a neural network based approach; but not clear that you couldn’t identify substructure classes better with a pure bioinformatic approach.
- wolverine876 3y agoThe OP says something different: "The insight here was that we could see what was being learned by the models to make their predictions that certain molecules would make for good antibiotics," James Collins, professor of Medical Engineering and Science at the Massachusetts Institute of Technology (MIT) and one of the study’s authors, said in a statement. ... "What we set out to do in this study was to open the black box. These models consist of very large numbers of calculations that mimic neural connections, and no one really knows what's going on underneath the hood," said Felix Wong, a postdoc at MIT and Harvard and one of the study’s lead authors. > not clear that you couldn’t identify substructure classes better with a pure bioinformatic approach. Lots of things aren't clear, but what supports your particular claim? Has it ever been done? And is it addressed in the article?
- hedora 3y agoThey’re saying that traditional computational drug discovery pipelines do all the stuff the authors are claiming. So, yes, it’s been done at scale for decades. The article makes all sorts of incorrect claims. For example, it says it has been 60 years since the last antibiotic was discovered, and that this works enables computational screening of antibiotics for the first time. Here’s work from 2022 that used conventional computational screening to discover a novel antibiotic: https://phys.org/news/2022-10-discovery-antibiotic-resistant-pathogens.html https://phys.org/news/2022-10-discovery-antibiotic-resistant...
- wolverine876 3y ago
- hoc 3y agoI would have expexted a more generative approach than the one they present. While this speeds up the screning process quite a lot you might as well feel like someone might just have thrown something out that might actually turn out useful and you can't backcheck without resorting to testing everthing. When I read the headline I was expected something more along "a network that knows what processes we have and therfor could imagine what compounds we could come up with" combined with that "inverse" network that predicts weaknesses of bacteria. Then those two would be matched and we would have to figure out the missing steps how to actually build the compound coming from that "black box" prediction. The benefit would be that we would be presented with something that could exist but we hadn't thought of because of our processes and limited resources. While I value the speedup and cost effect I'd think we could do much better here in exploiting the "creative" and powerful numerical inverse potential of ANNs especially in this area and outsmart those evolutionary skills of bacteria that way.
- stainablesteel 3y agoyour approach is a lot more complex, but they satisfy a lot of optimal criteria with a simpler approach, which i actually think is more impressive and also a much better target to hit. in fact i think they could have even achieved the same results using a much simpler exploratory method. a lot of biology, as well as where its been exploited, relies on specific binding, along with this comes the hope that what you test won't inadvertently show off-target effects. Through looking at what's already available for information you can actually use the lack something existing within a human system as a probabilistic edge over the future success of their application. in a regard its like an artistic rendering that makes use of negative space, its an impressive direction.
- anonuser123456 3y agoQuestion for biologists. How conserved are the genes that overcome antibiotics? If we stopped using them would bacteria lose the genes that provide resistance over time or do they gain that information forever?