Fig. 1

Overview of the construction process for RNAPSEC and prediction models. We first constructed RNAPSEC (A) and subsequently performed the model construction (B). A Overview of the dataset construction. The experimental information from RNAPhaSep, where multiple experiments were grouped together, was disassembled into individual experiments. To achieve this, we manually extracted the experimental information from the public literature stored in RNAPhaSep and recorded each experiment as one entry. B Overview of the model construction. We constructed two ML-based models using RNAPSEC. The first model predicts LLPS behavior using protein-derived features, RNA-derived features, and experimental conditions. The second model predicts the required conditions for a protein and RNA to undergo LLPS using protein-derived features and RNA-derived features. Abbreviations: conc., concentration