Find open-source science resources

piRNAs (short for PIWI-interacting RNAs) and their PIWI protein partners play a key role in fertility and maintaining genome integrity by restricting mobile genetic elements (transposons) in germ cells. piRNAs originate from genomic regions known as piRNA clusters. The piRNA Cluster Builder (PICB) is a versatile toolkit designed to identify genomic regions with a high density of piRNAs. It constructs piRNA clusters through a stepwise integration of unique and multimapping piRNAs and offers wide-ranging parameter settings, supported by an optimization function that allows users to test different parameter combinations to tailor the analysis to their specific piRNA system. The output includes extensive metadata columns, enabling researchers to rank clusters and extract cluster characteristics.

Coralysis is an R package featuring a multi-level integration algorithm for sensitive integration, reference-mapping, and cell-state identification in single-cell data. The multi-level integration algorithm is inspired by the process of assembling a puzzle - where one begins by grouping pieces based on low-to high-level features, such as color and shading, before looking into shape and patterns. This approach progressively blends the batch effects and separates cell types across multiple rounds of divisive clustering.

Public release of Profluent's ProGen3 protein language model family, including PMC-15B supporting sequence- and structure-conditioned generation for protein design, zero-shot fitness prediction, and antibody engineering with state-of-the-art performance on fitness and docking benchmarks (114+ stars, Apache 2.0)

topGO package provides tools for testing GO terms while accounting for the topology of the GO graph. Different test statistics and different methods for eliminating local similarities and dependencies between GO terms can be implemented and applied.

DeepSeek's open-source large language model for formal theorem proving in Lean 4, integrating informal and formal mathematical reasoning through recursive subgoal decomposition and reinforcement learning powered by DeepSeek-V3, with open weights and ProverBench evaluation (2025)

Easily get SRA download links and other information.

Most human genes have multiple promoters that control the expression of different isoforms. The use of these alternative promoters enables the regulation of isoform expression pre-transcriptionally. Alternative promoters have been found to be important in a wide number of cell types and diseases. proActiv is an R package that enables the analysis of promoters from RNA-seq data. proActiv uses aligned reads as input, and generates counts and normalized promoter activity estimates for each annotated promoter. In particular, proActiv accepts junction files from TopHat2 or STAR or BAM files as inputs. These estimates can then be used to identify which promoter is active, which promoter is inactive, and which promoters change their activity across conditions. proActiv also allows visualization of promoter activity across conditions.

High-throughput single-cell measurements of DNA methylation allows studying inter-cellular epigenetic heterogeneity, but this task faces the challenges of sparsity and noise. We present vmrseq, a statistical method that overcomes these challenges and identifies variably methylated regions accurately and robustly.

Therapeutics Data Commons: 66 AI-ready datasets across 22 drug discovery tasks with 29 leaderboards, covering target identification, molecular generation, ADMET prediction, and clinical trial outcomes (Harvard MIMS, NeurIPS 2021/2024)

A library for estimating thermochemical properties of molecules and adsorbates using group additivity.

Segment Anything in 3D medical images and videos, extending SAM2 to volumetric and temporal medical imaging with state-of-the-art zero-shot segmentation performance across CT, MRI, and surgical video (arXiv 2025)

Plots protein properties and visualizes position of peptide immunogens within protein sequence. Allows evaluation of immunogens based on structural and functional annotations to infer suitability for antibody-based methods aiming to detect native proteins.

AI agent for biological discovery and research automation

The DNEA R package is the latest implementation of the Differential Network Enrichment Analysis algorithm and is the successor to the Filigree Java-application described in Iyer et al. (2020). The package is designed to take as input an m x n expression matrix for some -omics modality (ie. metabolomics, lipidomics, proteomics, etc.) and jointly estimate the biological network associations of each condition using the DNEA algorithm described in Ma et al. (2019). This approach provides a framework for data-driven enrichment analysis across two experimental conditions that utilizes the underlying correlation structure of the data to determine feature-feature interactions.

Use hail via basilisk when appropriate, or via reticulate. This package can be used in terra.bio to interact with UK Biobank resources processed by hail.is.

saseR is a highly performant and fast framework for aberrant expression and splicing analyses. The main functions are: \itemize{ \item \code{\link{BamtoAspliCounts}} - Process BAM files to ASpli counts \item \code{\link{convertASpli}} - Get gene, bin or junction counts from ASpli SummarizedExperiment \item \code{\link{calculateOffsets}} - Create an offsets assays for aberrant expression or splicing analysis \item \code{\link{saseRfindEncodingDim}} - Estimate the optimal number of latent factors to include when estimating the mean expression \item \code{\link{saseRfit}} - Parameter estimation of the negative binomial distribution and compute p-values for aberrant expression and splicing } For information upon how to use these functions, check out our vignette at \url{https://github.com/statOmics/saseR/blob/main/vignettes/Vignette.Rmd} and the saseR paper: Segers, A. et al. (2023). Juggling offsets unlocks RNA-seq tools for fast scalable differential usage, aberrant splicing and expression analyses. bioRxiv. \url{https://doi.org/10.1101/2023.06.29.547014}.

First real quadrotor robot trained end-to-end with differentiable physics for vision-based agile flight, bridging simulation-based learning and real-world deployment with physics-informed neural network controllers (558+ stars)

Curated list of large weather models for AI Earth science

Curated scientific LLM papers (260+ models)

In genomics, differential analysis enables the discovery of groups of genes implicating important biological processes such as cell differentiation and aging. Non-parametric tests of differential gene expression usually detect shifts in centrality (such as mean or median), and therefore suffer from diminished power against alternative hypotheses characterized by shifts in spread (such as variance). This package provides a flexible family of non-parametric two-sample tests and K-sample tests, which is based on theoretical work around non-parametric tests, spacing statistics and local asymptotic normality (Erdmann-Pham et al., 2022+ [arXiv:2008.06664v2]; Erdmann-Pham, 2023+ [arXiv:2209.14235v2]).

This package is a Shiny app for interactively analyzing and visualizing Nanostring GeoMX Whole Transcriptome Atlas data. Users have the option of exploring a sample data to explore this app's functionality. Regions of interest (ROIs) can be filtered based on any user-provided metadata. Upon taking two or more groups of interest, all pairwise and ANOVA-like testing are automatically performed. Available ouputs include PCA, Volcano plots, tables and heatmaps. Aesthetics of each output are highly customizable.

Tools for plotting SingleCellExperiment objects in the chevreulPlot package. Includes functions for analysis and visualization of single-cell data. Supported by NIH grants R01CA137124 and R01EY026661 to David Cobrinik.

From the perspective of metabolites as the continuation of the central dogma of biology, metabolomics provides the closest link to many phenotypes of interest. This makes metabolomics research promising in teasing apart the complexities of living systems. However, due to experimental reasons, the data includes non-biological variation which limits quality and reproducibility, especially if the data is obtained from several batches. The batchCorr package reduces unwanted variation by way of between-batch alignment, within-batch drift correction and between-batch normalization using batch-specific quality control samples and long-term reference QC samples. Please see the associated article for more thorough descriptions of algorithms.

Docling-powered parsing with UI/CLI demonstration for rapid prototyping

Spatial homogeneous regions (SHRs) in tissues are domains that are homogenous with respect to cell type composition. We present a method for identifying SHRs using spatial transcriptomics data, and demonstrate that it is efficient and effective at finding SHRs for a wide variety of tissue types. concordex relies on analysis of k-nearest-neighbor (kNN) graphs. The tool is also useful for analysis of non-spatial transcriptomics data, and can elucidate the extent of concordance between partitions of cells derived from clustering algorithms, and transcriptomic similarity as represented in kNN graphs.

Digital Expression Explorer 2 (or DEE2 for short) is a repository of processed RNA-seq data in the form of counts. It was designed so that researchers could undertake re-analysis and meta-analysis of published RNA-seq studies quickly and easily. As of April 2020, over 1 million SRA datasets have been processed. This package provides an R interface to access these expression data. More information about the DEE2 project can be found at the project homepage (http://dee2.io) and main publication (https://doi.org/10.1093/gigascience/giz022).

In silico directed evolution framework using few-shot active learning to optimize protein activities, enabling rapid protein engineering with minimal experimental data (352+ stars, 2023)

This package implements SCnorm — a method to normalize single-cell RNA-seq data.

Extensible chemistry toolkit for MCP-enabled AI assistants, exposing molecule analysis, property prediction, and reaction synthesis tools through unified Python/MCP interfaces for chemistry agents and research workflows (Apache 2.0, 2025)

TOP constructs a transferable model across gene expression platforms for prospective experiments. Such a transferable model can be trained to make predictions on independent validation data with an accuracy that is similar to a re-substituted model. The TOP procedure also has the flexibility to be adapted to suit the most common clinical response variables, including linear response, binomial and Cox PH models.

animalcules is an R package for utilizing up-to-date data analytics, visualization methods, and machine learning models to provide users an easy-to-use interactive microbiome analysis framework. It can be used as a standalone software package or users can explore their data with the accompanying interactive R Shiny application. Traditional microbiome analysis such as alpha/beta diversity and differential abundance analysis are enhanced, while new methods like biomarker identification are introduced by animalcules. Powerful interactive and dynamic figures generated by animalcules enable users to understand their data better and discover new insights.

This package helps user to do easily RNA-seq data analysis with multiple methods (usually which needs many different input formats). Here the user will provid the expression data as a SummarizedExperiment object and will get results from different methods. It will help user to quickly evaluate different methods.

Tools for managing SingleCellExperiment objects as projects. Includes functions for analysis and visualization of single-cell data. Also included is a shiny app for visualization of pre-processed scRNA data. Supported by NIH grants R01CA137124 and R01EY026661 to David Cobrinik.

Large language-and-vision assistant for biomedicine, instruction-tuned on GPT-4-generated biomedical multimodal instruction-following data to enable conversational visual question answering over radiology, pathology, and microscopy images, establishing open recipes for adapting general vision-language models to the biomedical domain (Microsoft Research & University of Washington, 2.2K+ stars)

Toolkit for processing sequences in FASTA/Q formats.

Multimodal LLM for scientific charts and diagrams understanding/generation

A circos representation of multiple GWAS results.

Universal 3D molecular pretraining framework with 209M conformations, scaling to 1.1B parameters (Uni-Mol2) on 800M conformations for molecular property prediction, docking, and quantum chemistry (ICLR 2023, NeurIPS 2024)

End-to-end RNA 3D structure prediction using RNA language model pretrained on 23.7M sequences, outperforming existing methods and human expert groups on RNA-Puzzles and CASP15 (Nature Methods 2024)

Defines an S4 class for storing data from spatial -omics experiments. The class extends SingleCellExperiment to support storage and retrieval of additional information from spot-based and molecule-based platforms, including spatial coordinates, images, and image metadata. A specialized constructor function is included for data from the 10x Genomics Visium platform.

G-quadruplexes (G4s) are unique nucleic acid secondary structures predominantly found in guanine-rich regions and have been shown to be involved in various biological regulatory processes. G4SNVHunter is an R package designed to rapidly identify genomic sequences with G4-forming propensity and to accurately screen user-provided single nucleotide variants—as well as other small-scale variants such as indels and MNVs—for their potential to destabilize these structures. This allows researchers to then screen these critical variants for deeper study, digging into how they might influence biological functions—think gene regulation, for instance—by impairing G4 formation propensity.

Advanced paper search agent powered by large language models, autonomously invoking search tools, reading papers, and selecting references to deliver comprehensive and accurate results for complex scholarly queries (1.5K+ stars, Apache 2.0, 2024)

RNA foundation model trained on millions of RNA sequences for generalist RNA sequence understanding, enabling downstream structure prediction, function annotation, and representation learning for non-coding RNAs (ml4bio, 372+ stars)

Provides SummarizedExperiment-like containers for storing and manipulating dimensionally-reduced assay data. The ReducedExperiment classes allow users to simultaneously manipulate their original dataset and their decomposed data, in addition to other method-specific outputs like feature loadings. Implements utilities and specialised classes for the application of stabilised independent component analysis (sICA) and weighted gene correlation network analysis (WGCNA).

State-of-the-art pretrained language models for proteins trained on thousands of GPUs and Google TPUs using Transformer architectures, enabling protein property prediction, feature extraction, and transfer learning across diverse downstream tasks (1.3K+ stars, MIT, 2020-2026)

GRIDSS: the Genomic Rearrangement IDentification Software Suite.

General purpose tools for high-throughput catalysis.

systemPipeShiny (SPS) extends the widely used systemPipeR (SPR) workflow environment with a versatile graphical user interface provided by a Shiny App. This allows non-R users, such as experimentalists, to run many systemPipeR’s workflow designs, control, and visualization functionalities interactively without requiring knowledge of R. Most importantly, SPS has been designed as a general purpose framework for interacting with other R packages in an intuitive manner. Like most Shiny Apps, SPS can be used on both local computers as well as centralized server-based deployments that can be accessed remotely as a public web service for using SPR’s functionalities with community and/or private data. The framework can integrate many core packages from the R/Bioconductor ecosystem. Examples of SPS’ current functionalities include: (a) interactive creation of experimental designs and metadata using an easy to use tabular editor or file uploader; (b) visualization of workflow topologies combined with auto-generation of R Markdown preview for interactively designed workflows; (d) access to a wide range of data processing routines; (e) and an extendable set of visualization functionalities. Complex visual results can be managed on a 'Canvas Workbench’ allowing users to organize and to compare plots in an efficient manner combined with a session snapshot feature to continue work at a later time. The present suite of pre-configured visualization examples. The modular design of SPR makes it easy to design custom functions without any knowledge of Shiny, as well as extending the environment in the future with contributions from the community.

Collection of tools for working with BAM files.