🧠 The complete model and datasets are available open source:

Trained on general AMP datasets and fine-tuned on TB-specific sequences, the best-performing model achieved 90% accuracy and an AUC of 0.97, while generating 94 novel peptides predicted to have antimicrobial properties.

The study presents an LSTM-based #deeplearning framework combining classification and generative modeling to identify peptides active against Mycobacterium tuberculosis.

It examines how CNNs, RNNs, GNNs, and transformer architectures have improved predictions for transcription factor binding, chromatin accessibility, splicing, and other regulatory tasks—highlighting breakthroughs, limitations, and interpretability methods that enhance biological understanding.

This comprehensive survey reviews recent advances in applying #deeplearning to #generegulation research.

🧠 CoRTE is open source and freely available on GitHub: 

Implemented in #Python with GTEx integration, it identifies time-dependent gene interactions and supports visualization through Cytoscape or TimeNexus for exploring aging-related and disease-associated transcriptional dynamics.

CoRTE constructs temporal #gene co-expression networks (TGCNs) from genotype-tissue expression data using statistical correlation across age-defined time points.

⚙️ Source code and datasets available #opensource:

Tested on drug–drug interaction and drug–disease association datasets, it achieved higher or comparable accuracy and F-scores to leading models while maintaining efficiency and interpretability. The method generalizes across diverse biological prediction tasks.

PSO-FeatureFusion integrates particle swarm optimization with #neuralnetworks to dynamically combine heterogeneous biological features.

📊 Access Shiny-Calorie online: https://shiny.iaas.uni-bonn.de/Shiny-Calorie 
Source code and API reference are available here:

It harmonizes metadata, automates quality control, supports multi-cohort studies, and provides advanced statistical tools for energy expenditure and metabolism analyses with publication-ready visualizations.

Shiny-Calorie is an open-source R-based platform for integrating, analyzing, and visualizing indirect calorimetry data across multiple metabolic phenotyping systems.

🧰 Source code and datasets are openly available:

Validated on DrugCombDB and O’Neil datasets, the model outperformed existing methods, achieving AUC-ROC above 0.98 and identifying eight potential novel synergistic drug combinations.

DSA-DeepFM integrates a dual-stage attention mechanism with factorization machines to predict synergistic anticancer drug pairs by modeling both categorical and numerical biological features.

💾 Open-source code for tidyGWAS and LDSC is available here: 
🔸https://github.com/Ararder/tidyGWAS 
🔸https://github.com/Ararder/ldsR  

Docker container:

It harmonizes variant identifiers across genome builds, imputes missing columns, and outputs partitioned parquet files. Benchmarks show up to 6.5x faster processing and greater memory efficiency than existing tools.