{
  "id": 17234689,
  "title": "Degree bias predicts, outperforms and improves machine learning phage-host interaction predictions",
  "url": "https://urgent.news/2026/10/30/degree-bias-predicts-outperforms-and-improves-machine-learning-phage",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-30T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.10.29.762857v1?rss=1"
  },
  "original_language": "en",
  "account": "Machine learning (ML) models are frequently utilized to predict phage-host interactions and guide phage therapy. The Area Under the Receiver Operating Characteristic (AUROC) metric is the standard for evaluating these models, with reported high scores ranging from 0.7 to 0.95, indicating strong predictive power. However, the highly structured phage-bacterium networks can lead to inflated scores due to degree bias, which is the baseline likelihood of infection driven purely by the average phage infectivity and bacterial susceptibility. In this study, researchers found that an interaction-naive model based solely on network degrees outperforms five published ML models across ten datasets, suggesting that global AUROC scores primarily reflect network structure rather than true interactions. By using randomized degree-conserved networks and degree-stratified evaluation, the researchers were able to isolate true interaction predictions and demonstrate that current ML model performance is significantly lower than the global AUROC suggests. Moreover, the study reveals that removing degree bias from training data improves interaction predictability and enhances both global and stratified performance. This work provides a crucial framework for model evaluation and offers a simple method to enhance ML-guided phage selection.",
  "summary": "Machine learning (ML) models predicting phage-host interactions are widely used to prioritize phages for phage therapy. The Area Under the Receiver Operating Characteristic (AUROC) metric is standard for model evaluation, with reported high scores (0.7 - 0.95) suggesting strong predictive power. However, because phage-bacterium networks are highly structured, these scores can be inflated by…",
  "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."
}