Find open-source science resources

KIMMDY is a biomolecular reaction emulator that simulates complex biochemical reactions across conformational ensembles using kinetic Monte Carlo. It leverages graph neural networks to model dynamic, large-scale systems involving successive, competing reactions in biomolecules like proteins and DNA.

RiSPICE (Rice SNP Prioritization Integrating Chromatin Effects) is a computational framework for prioritizing non-coding rice variants by integrating predicted chromatin effects from a fine-tuned DNA language model.

aims-PAX is a high-performance, modular active learning framework designed to expedite the development of machine learning force fields (MLFFs) for complex molecular and material simulations.

PNcsp+ is an enhanced, chemically interpretable framework for Crystal Structure Prediction (CSP) that leverages the Mendeleev Periodic Number (PN) to guide elemental similarity. It combines state-of-the-art neural networks (MACE, M3GNet, ALIGNN-FF) for accelerated, accurate prediction of crystal structures in both inorganic and hybrid chemical spaces.

M2UMol is a multimodal pre-training framework designed for molecular representation learning, enabling accurate predictions even when incomplete molecular data modalities are provided.

crisprHAL is a deep learning model developed for predicting on-target Cas9 cleavage activity using single guide RNA (sgRNA) sequences. It improves prediction performance through data-centric methods, such as optimizing input length and rigorous data filtering.

EquiHGNN is a framework that uses equivariant hypergraph neural networks to model complex molecular systems. It enables multi-way interactions beyond traditional pairwise graphs for improved molecular representation.

MIL-Adapter is a novel framework designed for few-shot slide-level classification in computational pathology. It combines Multiple Instance Learning (MIL) aggregation functions with lightweight vision-language adapters to process Whole Slide Images (WSIs) for cancer diagnosis.

scRevol is an RL-based model designed to infer tumor evolution and characterize clonal structure from single-cell RNA sequencing (scRNA-seq) data. It utilizes CNV profiles to reconstruct evolutionary trajectories and resolve subclonal diversity in cancer.

StabLyzeGraph is a computational framework utilizing Graph Neural Networks to analyze and predict stabilizing mutations in proteins, accelerating rational protein design. It enables high-throughput screening of combinatorial mutations by integrating structural, physicochemical, and evolutionary data.

Biomed-DPT is a knowledge-enhanced framework for adapting pre-trained Vision-Language Models (VLMs) to biomedical image classification tasks. It uses a dual prompt tuning mechanism incorporating ensemble clinical prompts and LLM-driven expert prompts.

Viral Host UnveiLing Kit - A tool-kit for phage host prediction. This tool is for bacteriophage host prediction based on annotated genomic features and deep neural networks.

MedSlice is an open-source pipeline of fine-tuned open-source LLMs (best: Llama 3.1 8B) that automatically extracts History of Present Illness, Interval History, and Assessment and Plan sections from clinical progress notes, outperforming GPT-4o (F1 0.92) while keeping data local for privacy.

DeepDILI is a tool for Deep Learning-Powered Drug-Induced Liver Injury Prediction Using Model-Level Representation

A multimodal deep learning model, DeepNCI, is proposed for improving noncovalent interactions (NCIs) calculated via density functional theory (DFT).

interpretable sequence-based prediction of type IV secreted effectors using eXtreme gradient boosting algorithm. Type IV secreted effectors (T4SEs) can be translocated into the cytosol of host cells via type IV secretion system (T4SS) and cause diseases.

Predicting the effect of mutations on protein-RNA binding with Deep Learning | This repository contains all DeepCLIP Python code | A context-aware neural network for modeling and predicting protein binding to nucleic acids using only sequence input | DeepCLIP is a neural network with shallow convolutional layers connected to a bidirectional LSTM layer

The following tool computes background parenchyma enhancement quantification on DCE-MRI at baseline and first time point with enhancement.

A framework-agnostic Python toolkit that provides structured and reliable cross-validation methods (including grouped, temporal, and spatial splits) for mitigating data leakage in medical machine learning applications.

A framework-agnostic Python toolkit for robust cross-validation in medical machine learning. It supports numerous CV methods, detects data leakage, and provides structure-aware evaluation for biomedical data.

graphpancake is an open-source Python package that generates molecular graphs for small organic molecules by translating quantum mechanical data (DFT/WFT) into graph representations. It is designed for cheminformatics pipelines, enabling advanced predictive modeling like QSAR.

RS-WaterQuality Mapper is an open-source QGIS plugin and scientific workflow designed for water quality analysis using satellite remote sensing data. It applies advanced machine learning techniques, such as multi-predictor ensemble models, to enhance predictive accuracy in aquatic environments.

Convpaint is a framework that performs interactive pixel classification/image segmentation by leveraging pretrained deep learning networks (CNNs, ViTs) and fast machine learning classifiers. It operates within the napari ecosystem, supporting image analysis workflows from subcellular to animal scale.

Provides an R wrapper for the implementation of FI-tSNE from the python package openTNSE. See Poličar et al. (2019) <doi:10.1101/731877> and the algorithm described by Linderman et al. (2018) <doi:10.1038/s41592-018-0308-4>.

A weighted ensemble convolutional neural network for the virulence prediction of influenza A virus using all 8 segments. Influenza viruses are persistently threatening public health, causing annual epidemics and sporadic pandemics. The evolution of influenza viruses remains to be the main obstacle in the effectiveness of antiviral treatments due to rapid mutations. The goal of this work is to predict the virulence of influenza viruses using sequence data.

PyMVPD is a Python-based toolbox to model the multivariate interactions between brain regions using fMRI data.

iROAR (Immune Repertoire Over Amplification) is a tool for multiplex PCR-specific bias evaluation and correction to adjust clonal count of VDJtool output table.

Ethoflow is a computer vision and artificial intelligence-based software for automatic behavior analysis.

DeepEthogram is a Python package for automatically classifying each frame of a video into a set of pre-defined behaviors. Designed for neuroscience research, it could be used in any scenario where you need to detect actions from each frame of a video.

CoNSEPT (Convolutional Neural Network-based Sequence-to-Expression Prediction Tool) is a tool to predict gene expression in various cis and trans contexts. Inputs to CoNSEPT are enhancer sequence, transcription factor levels in one or many trans conditions, TF motifs (PWMs), and any prior knowledge of TF-TF interactions.

Coevolution and Convolutional Neural Network (CoCoNet) is a meyhod used for RNA contact prediction.

Accurate Clustering-based Ensemble Method to Predict Motor Imagery Tasks from Multi-channel EEG Data.

ClaMSA is tool that uses machine learning to classify sequences that are related to each other via a tree. It takes as input a tree and a multiple sequence alignment (MSA) of sequences and outputs probabilities that the MSA belongs to given classes.

An active learning platform for expert-guided, data efficient discovery of behavior

An in-silico method for predicting the thermal activity for xylanases from GH10 and GH11 families. a multi-step method for the classification of xylanases from glycoside hydrolases families 10 and 11 based on their optimum temperature of activity. TAXyl: An in-silico method for predicting the optimum temperature of activity for xylanases from GH10 and GH11 families. Xylanases are involved in lignocellulosic biomass degradation and the enzyme optimum temperature is an essential factor to be considered for choosing an appropriate biocatalyst for specific usage

> HOMEPAGE MISSING! | > CORRECT NAME OF TOOL COULD ALSO BE 'hyperspectral', 'vegetation', 'planation', 'Cunninghamia' | UAV-Based High-Throughput Approach for Fast Growing <em>Cunninghamia lanceolata (Lamb.)</em> Cultivar Screening by Machine Learning | Accurate measurements of tree height and diameter at breast height (DBH) in forests to evaluate the growth rate of cultivars is still a significant challenge, even when using LiDAR and 3-D modeling. We propose an integrated pipeline methodology to measure the biomass of different tree cultivars in plantation forests with high crown density which that combines unmanned aerial vehicles (UAVs), hyperspectral image sensors, and data processing algorithms using machine learning. Using a planation of Cunninghamia lanceolate, commonly known as Chinese fir, in Fujian, China, images were collected using a hyperspectral camera and orthorectified in HiSpectral Stitcher

Bipartite Tight Spectral Clustering (BiTSC) Algorithm for Identifying Conserved Gene Co-clusters in Two Species. Yidan Eden Sun, Heather J. Zhou and Jingyi Jessica Li. $ python3 BiTSC.py ./data/orthologs_data_uniq.csv ./data/dm_timecourse_FPKMs.csv ./data/ce_timecourse_FPKMs.csv 30 10 100 0.8 10 10 ./data/result 0.90 0.95 1.00

using deep learning to classify riboswitches with ~99% accuracy. Classifying Putative Riboswitch Sequences. RIBOFLOW - classifying riboswitches with >99% accuracy. riboflow is a python package for classifying putative riboswitch sequences into one of 32 classes with > 99% accuracy. It is based on a tensorflow deep learning model. riboflow has been tested using Python 3.5.2

A Graph-Based Deep Learning Framework for Chemical Structures. kGCN: a graph-based deep learning framework for life science.

Joint Integration and Discrimination for Automated Single-Cell Annotation. JIND is a framework for automated cell-type identification based on neural networks. It directly learns a low-dimensional representation (latent code) inwhich cell-types can be reliably determined. To account for batch effects, JIND performs a novel asymmetric alignment in which the transcriptomic profileof unseen cells is mapped onto the previously learned latent space, hence avoiding the need of retraining the model whenever a new dataset becomes available. JIND also learns cell-type-specific confidence thresholds to identify and reject cells that cannot be reliably classified. We show on datasets with and without batch effects that JIND classifies cells more accurately than previously proposed methods while rejecting only a small proportion of cells. Moreover, JIND batch alignment is parallelizable, being more than five or six times faster than Seurat integration.

Gilda is a Python package and REST service that grounds (i.e., finds appropriate identifiers in namespaces for) named entities in biomedical text.

FastISM is a Python package for in-silico saturation mutagenesis for convolutional neural networks.

A sequence-based solubility predictor created with dilated squeeze excitation residual networks.

DeepHE is a deep learning framework for essential gene prediction based on sequence data and PPI network.

DeePaC-Live is a deep learning framework for real-time detection of novel pathogens during sequencing.

comboLTR is a new polynomial regression-based framework for modeling anti-cancer effects of drug combinations in various doses and across different cancer cell lines.

COmpositional Data Analysis with Recursive Feature Elimination.

An Open Source Unsupervised Algorithm for Discovery of Spontaneous Behaviors | Behavioral segmentation of open field in DeepLabCut, or B-SOID ("B-side"), is an unsupervised learning algorithm written in MATLAB that serves to discover behaviors that are not pre-defined by users | B-SOID: Behavioral segmentation of open field in DeepLabCut

Multi-task self-supervised learning for cell type classification. Rapidly generated scRNA-seq datasets enable us to understand cellular differences and the function of each individual cell at single-cell resolution.