🧰 Open-source code available at  https://github.com/opencobra/cobratoolbox. The toolbox is implemented in MATLAB within the COBRA Toolbox framework and includes template files for dietary data input and database food equivalent selection.

Using keyword matching & Euclidean distance calculations, it pairs dietary items with nearly 15,400 database entries & converts macronutrient data to mmol/day/person. When integrated with the nutrition algorithm, it calculates minimal dietary changes to achieve targeted metabolic flux modifications.

The Nutrition Toolbox processes USDA and FRIDA databases to convert dietary food items into metabolite-level constraints for whole-body metabolic models.

🐍 Open-source pipeline available at  https://github.com/itvgenomics/pipeasm. 
The pipeline is implemented in Python/Snakemake with Singularity and runs on Unix-based systems.

Applied to 4 vertebrate species with distinct genomic characteristics, it produced assemblies with scaffold L50 proportional to expected chromosome length and up to 99.6% BUSCO completeness from a single command and configuration file.

Pipeasm is a Snakemake-based genome assembly pipeline containerized with Singularity that automates read trimming, nuclear and mitogenome assembly, scaffolding, decontamination, and quality evaluation using HiFi, ONT, and Hi-C data.

🌐 Web application available at https://ggplotagent.databio1.com/ with source code at https://github.com/charlin90/ggplotAgent under MIT License. An offline Streamlit app is also available for local deployment.

Benchmarked against DeepSeek-V3 on 20 visualization tasks, it achieved 100% code executability vs 85% and a Publication-Ready Score of 1.9 versus 0.7. The system uses specialized agents for validation, planning, code generation, execution, and visual checking to iteratively debug and refine outputs.

ggplotAgent is a multi-modal AI agent that automates publication-ready ggplot2 visualizations using a dual-layered debugging framework.

🦀 Open-source Rust library available at https://github.com/microscopic-image-analysis/zelll and https://crates.io/crates/zelll under MIT license. 

Python bindings are available at https://pypi.org/project/zelll for core functionality.

Benchmarked on synthetic point cloud data, the implementation achieves O(n) time complexity with iteration agnostic to data structure. A case study demonstrates protein surface sampling using smooth distance functions with individual queries completing in microseconds.

zelll implements the cell lists algorithm in Rust using sparse storage for partitioning grids, enabling dynamic simulation boundaries and sparsely populated spaces for coarse-grained applications like chromosome structure modeling.

💻 Code and models at https://github.com/emiliostassios/Classification-of-divergent-genes-using-ML.

Applied to over 600k gut microbiome orphans, ~30% were predicted divergent and showed significantly higher intrinsic disorder and shorter length than others.

Machine learning classifiers trained on simulated diverged orphan sequences achieved ~90% accuracy under moderate divergence and ~70% under extreme conditions using features from non-significant DIAMOND hits.

💻 REDAC is available at https://github.com/franruss/REDAC and online at https://frusso.shinyapps.io/REDAC. 

The platform includes a PubMed-based RAG module for biological interpretation, edgeR-based workflows, and supports 16 pathway enrichment databases with complete code transparency.

JSON-structured outputs reduced hallucinations from 89% to 24.7% in 150 validation tests. Applied to Gefitinib-resistant lung cancer cells, REDAC identified key pathways including TNF and MAPK signaling, generating automated Word and HTML reports with interactive visualizations.

REDAC is a web-based R application using Gemma and LLaMA LLMs for differential expression analysis via natural language queries.

💻 Open-source software available at https://github.com/Josh-Brand/Fluoro-forest under the MIT license. The pipeline includes cluster-based sampling, a multichannel annotation interface, and outputs cell type probabilities with performance metrics.