{
  "id": 9885673,
  "title": "Predictors Recover Most of the Metagenomic Signal for Antibiotic Resistance Gene Occurrence: A Cross-City Test of Geographic Transferability in Urban Wastewater",
  "url": "https://urgent.news/2026/09/25/predictors-recover-most-of-the-metagenomic-signal-for-antibiotic",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-25T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.24.753270v1?rss=1"
  },
  "original_language": "en",
  "account": "Antibiotic resistance genes (ARGs) spread through cities into rivers and coastal waters via wastewater treatment plants. Monitoring ARGs typically necessitates metagenomic sequencing, a process that proves prohibitively expensive, leading most utilities to sample only sporadically. Fortunately, weather and location data are readily accessible on a daily basis for virtually any location worldwide, rendering them appealing predictors to guide the allocation of limited sequencing resources. However, the extent to which these models can generalize to cities not previously included in the training data remains largely unclear, as most published studies evaluate them within the same catchments, which amounts to testing interpolation rather than geographic transferability.\n\nIn order to tackle this issue, we linked 235 wastewater metagenomes collected from five European cities with 23 abiotic predictors encompassing geospatial, meteorological, hydrological, radiative, and temporal domains. Model efficacy was assessed using leave-one-group-out (LOGO) cross-validation, wherein all samples from one city were excluded for testing while the remaining cities served as the training set. CatBoost demonstrated a median LOGO ROC-AUC of 0.929 and a median F1-score of 0.750. Utilizing freely available environmental reanalysis predictors, including meteorological, hydrological, radiative, geospatial, and temporal data, CatBoost achieved a mean ROC-AUC of 0.722, thereby recovering 78% of the predictive performance of the comprehensive omics-integrated model. Interestingly, the elimination of latitude and longitude resulted in a mere 0.003 reduction in ROC-AUC, whereas replacing random cross-validation with city-wise validation led to a 0.052 decrease. Predictive performance varied across ARG classes, spanning from a remarkable ROC-AUC of 0.981 for beta-lactam resistance genes to 0.762 for glycopeptide resistance genes. These results underscore that freely accessible environmental reanalysis predictors, spanning various domains, can recover 78% of the predictive signal for antibiotic resistance gene occurrence in urban wastewater. Consequently, scarce sequencing capacity can be strategically directed towards catchments where such decisions are likely to change.",
  "summary": "Antibiotic resistance genes (ARGs) travel from cities into rivers and coastal waters through wastewater treatment plants. Monitoring them normally requires metagenomic sequencing, which is sufficiently costly that most utilities can sample only occasionally. Weather and location data are freely available on a daily basis for virtually any location worldwide, making them attractive predictors for…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}