The article by Peng et al. addresses the challenge of predicting intrinsic protein disorder, particularly focusing on the length-dependent nature of disordered regions. The authors propose two new predictor models, VSL2-M1 and VSL2-M2, which are designed to improve the accuracy of predicting short (≤30 residues) and long (>30 residues) disordered regions. These models integrate specialized predictors for short and long disordered regions into a meta-predictor, achieving well-balanced prediction accuracies of 81% on both types of disordered regions. The VSL2 predictors outperform several existing predictors in terms of accuracy, sensitivity, and specificity. The study highlights the importance of considering the length-dependent amino acid compositions and sequence properties of disordered regions, and provides a freely accessible predictor for non-commercial use.The article by Peng et al. addresses the challenge of predicting intrinsic protein disorder, particularly focusing on the length-dependent nature of disordered regions. The authors propose two new predictor models, VSL2-M1 and VSL2-M2, which are designed to improve the accuracy of predicting short (≤30 residues) and long (>30 residues) disordered regions. These models integrate specialized predictors for short and long disordered regions into a meta-predictor, achieving well-balanced prediction accuracies of 81% on both types of disordered regions. The VSL2 predictors outperform several existing predictors in terms of accuracy, sensitivity, and specificity. The study highlights the importance of considering the length-dependent amino acid compositions and sequence properties of disordered regions, and provides a freely accessible predictor for non-commercial use.