AI summaryⓘ
The authors studied how to figure out which antibiotics are actively affecting bacteria, rather than just which ones were applied, since resistant bacteria won't show changes from some drugs. They found that usual methods either assume the applied drugs are active or struggle to identify the active ones correctly. To solve this, they created AURA, a new approach that narrows down the active antibiotics by looking at changes in bacteria's appearance and ensuring the active set is a subset of those applied. AURA worked much better than previous tools, correctly identifying the active antibiotics about 95% of the time across different experiments. Their method also includes a way to withhold unsure predictions for more reliability.
antibiotic resistanceE. colicytological profilinginverse attributionenergy-based modelsmorphological responsecross-replicate transferactive drug subsetstrain variabilityprediction abstention
Authors
Kartik Jhawar, Mrunmayee Deshpande, Wilfried Moreira, Guillermo C. Bazan, Lipo Wang
Abstract
When a bacterial sample is exposed to several antibiotics, not every applied drug necessarily acts: if the organism is resistant to one of them, that drug leaves no morphological trace. The clinically meaningful quantity is therefore not which antibiotics were applied, but which ones were active. We show that these two are sharply decoupled in real E. coli microscopy - naively assuming the applied combination equals the active one is correct only about 37% of the time - yet existing computational tools are ill-suited to recovering the active set. Forward perturbation models such as scGen, CPA, and IMPA are designed to predict appearance from treatment, not the reverse, and inverting them degrades sharply; discriminative image classifiers tend to memorise strain- and batch-specific texture and fail to transfer across experimental replicates. We introduce AURA, which reframes the task as constrained, energy-based inverse attribution. Its central inductive bias is that the active set must be a subset of the applied set; this collapses the candidate space and lets AURA infer the active subset of applied antibiotics by decomposing residual morphology into antibiotic response atoms and selecting the subset with the lowest reconstruction energy, using no strain label at test time. AURA-E adds evidence-aware abstention, withholding a prediction when candidate explanations remain near-equally plausible. On cross-replicate transfer in an E. coli cytological profiling dataset, AURA recovers the active antibiotic combination with 95.47% exact-match accuracy.