More Like this

1. PepBERT: Lightweight language models for bioactive peptide representation

   https://doi.org/10.1101/2025.04.08.647838  

   Du, Zhenjiao; Caragea, Doina; Guo, Xiaolong; Li, Yonghui (April 2025, bioRxiv)

   Abstract Protein language models (pLMs) have been widely adopted for various protein and peptide-related downstream tasks and demonstrated promising performance. However, short peptides are significantly underrepresented in commonly used pLM training datasets. For example, only 2.8% of sequences in the UniProt Reference Cluster (UniRef) contain fewer than 50 residues, which potentially limits the effectiveness of pLMs for peptide-specific applications. Here, we present PepBERT, a lightweight and efficient peptide language model specifically designed for encoding peptide sequences. Two versions of the model—PepBERT-large (4.9 million parameters) and PepBERT-small (1.86 million parameters)—were pretrained from scratch using four custom peptide datasets and evaluated on nine peptide-related downstream prediction tasks. Both PepBERT models achieved performance superior to or comparable to the benchmark model, ESM-2 with 7.5 million parameters, on 8 out of 9 datasets. Overall, PepBERT provides a compact yet effective solution for generating high-quality peptide representations for downstream applications. By enabling more accurate representation and prediction of bioactive peptides, PepBERT can accelerate the discovery of food-derived bioactive peptides with health-promoting properties, supporting the development of sustainable functional foods and value-added utilization of food processing by-products. The datasets, source codes, pretrained models, and tutorials for the usage of PepBERT are available athttps://github.com/dzjxzyd/PepBERT. 

   more » « less
2. Estimation of substitution and indel rates via k-mer statistics

   https://doi.org/10.1186/s13015-026-00299-9  

   Hera, Mahmudur Rahman; Medvedev, Paul; Koslicki, David; Blanca, Antonio (July 2026, Algorithms for Molecular Biology)

   Abstract Methods utilizing$$k$$ $k$ -mers are widely used in bioinformatics, yet our understanding of their statistical properties under realistic mutation models remains incomplete. Previously, substitution-only mutation models have been considered to derive precise expectations and variances for mutated$$k$$ $k$ -mers and intervals of mutated and non-mutated sequences. In this work, we consider a mutation model that incorporates insertions and deletions in addition to single-nucleotide substitutions. Within this framework, we derive closed-formk-mer-based estimators for the three fundamental mutation parameters: substitution, deletion rate, and insertion rates. We provide theoretical guarantees in the form of concentration inequalities, ensuring accuracy of our estimators under reasonable model assumptions. Empirical evaluations on simulated evolution of genomic sequences confirm our theoretical findings, demonstrating that accounting for insertions and deletions signals allows for accurate estimation of mutation rates and improves upon the results obtained by considering a substitution-only model. An implementation of estimating the mutation parameters from a pair of fasta files is available here:https://github.com/KoslickiLab/estimate_rates_using_mutation_model.git. The results presented in this manuscript can be reproduced using the code available here:https://github.com/KoslickiLab/est_rates_experiments.git. 

   more » « less
3. E(3) equivariant graph neural networks for robust and accurate protein-protein interaction site prediction

   https://doi.org/10.1371/journal.pcbi.1011435  

   Roche, Rahmatullah; Moussad, Bernard; Shuvo, Md Hossain; Bhattacharya, Debswapna (August 2023, PLOS Computational Biology)

   Li, Jinyan (Ed.)

   Artificial intelligence-powered protein structure prediction methods have led to a paradigm-shift in computational structural biology, yet contemporary approaches for predicting the interfacial residues (i.e., sites) of protein-protein interaction (PPI) still rely on experimental structures. Recent studies have demonstrated benefits of employing graph convolution for PPI site prediction, but ignore symmetries naturally occurring in 3-dimensional space and act only on experimental coordinates. Here we present EquiPPIS, an E(3) equivariant graph neural network approach for PPI site prediction. EquiPPIS employs symmetry-aware graph convolutions that transform equivariantly with translation, rotation, and reflection in 3D space, providing richer representations for molecular data compared to invariant convolutions. EquiPPIS substantially outperforms state-of-the-art approaches based on the same experimental input, and exhibits remarkable robustness by attaining better accuracy with predicted structural models from AlphaFold2 than what existing methods can achieve even with experimental structures. Freely available athttps://github.com/Bhattacharya-Lab/EquiPPIS, EquiPPIS enables accurate PPI site prediction at scale. 

   more » « less
4. PepCNN deep learning tool for predicting peptide binding residues in proteins using sequence, structural, and language model features

   https://doi.org/10.1038/s41598-023-47624-5  

   Chandra, Abel; Sharma, Alok; Dehzangi, Iman; Tsunoda, Tatsuhiko; Sattar, Abdul (December 2023, Scientific Reports)

   Abstract Protein–peptide interactions play a crucial role in various cellular processes and are implicated in abnormal cellular behaviors leading to diseases such as cancer. Therefore, understanding these interactions is vital for both functional genomics and drug discovery efforts. Despite a significant increase in the availability of protein–peptide complexes, experimental methods for studying these interactions remain laborious, time-consuming, and expensive. Computational methods offer a complementary approach but often fall short in terms of prediction accuracy. To address these challenges, we introduce PepCNN, a deep learning-based prediction model that incorporates structural and sequence-based information from primary protein sequences. By utilizing a combination of half-sphere exposure, position specific scoring matrices from multiple-sequence alignment tool, and embedding from a pre-trained protein language model, PepCNN outperforms state-of-the-art methods in terms of specificity, precision, and AUC. The PepCNN software and datasets are publicly available athttps://github.com/abelavit/PepCNN.git. 

   more » « less
5. Towards instance-wise calibration: local amortized diagnostics and reshaping of conditional densities (LADaR)

   https://doi.org/10.1088/2632-2153/ae1f05  

   Dey, Biprateep; Zhao, David; Andrews, Brett H; Newman, Jeffrey A; Izbicki, Rafael; Lee, Ann B (December 2025, Machine Learning: Science and Technology)

   Abstract Key science questions, such as galaxy distance estimation and weather forecasting, often require knowing the full predictive distribution of a target variableYgiven complex inputsX. Despite recent advances in machine learning and physics-based models, it remains challenging to assess whether an initial model is calibrated for allx, and when needed, to reshape the densities ofytoward ‘instance-wise’ calibration. This paper introduces the local amortized diagnostics and reshaping of conditional densities (LADaR) framework and proposes a new computationally efficient algorithm (Cal-PIT) that produces interpretable local diagnostics and provides a mechanism for adjusting conditional density estimates (CDEs).Cal-PITlearns a single interpretable local probability–probability map from calibration data that identifies where and how the initial model is miscalibrated across feature space, which can be used to morph CDEs such that they are well-calibrated. We illustrate the LADaR framework on synthetic examples, including probabilistic forecasting from image sequences, akin to predicting storm wind speed from satellite imagery. Our main science application involves estimating the probability density functions of galaxy distances given photometric data, whereCal-PITachieves better instance-wise calibration than all 11 other literature methods in a benchmark data challenge, demonstrating its utility for next-generation cosmological analyzes⁹9Code available as a Python package here:https://github.com/lee-group-cmu/Cal-PIT.. 

   more » « less