Multimodal Models for Integrated Regulatory Analysis (MIRA)
To Know
About this Class
Join us for a webinar introducing MIRA, a Python software package designed for the analysis of single-cell RNA-seq, ATAC-seq, and multimodal data.
Discover how MIRA helps decode the cis-regulatory logic of gene expression, including the development of mathematical models that integrate scATAC-seq and scRNA-seq data. MIRA also models latent cell states, reconstructs developmental trajectories, and quantifies how cis-regulatory activity influences transcription.
The webinar will also highlight:
- CODAL, which enables multi-batch analysis using a variational autoencoder with mutual information regularization to separate biological signals from batch effects in perturbation experiments.
- scEpiSparX, which represents genomic intervals as epigenetic embeddings to enable efficient deep learning models for regulatory element prediction.
Together, these tools offer powerful approaches for modeling cis-regulatory mechanisms across diverse contexts. NCI’s Informatics Technology for Cancer Research program funds this tool.