Wei Li Lab software
Computational tools for CRISPR screens, genome engineering, and transcriptomics.
This page links maintained applications and repositories when available. Older methods are retained as part of the lab’s computational record and link to their publications or archived code.
CRISPR screen analysis
From quality control to hit identification
Tools for pooled CRISPR-screen analysis, visualization, and integrative interpretation.
MAGeCK
Model-based identification of positively and negatively selected genes from genome-scale CRISPR-Cas9 knockout screens.
The ongoing MAGeCK2 project extends MAGeCK with more functionalities, including paired-sample analysis, UMI and paired-guide screens.
MAGeCKFlute
Integrative analysis and visualization of pooled CRISPR screens, including quality assessment and pathway-level interpretation.
MAGeCK-VISPR
A workflow for quality control, modeling, and visualization of genome-wide CRISPR screens.
CRISP-view
A published database of functional genetic screens spanning multiple phenotypes. The original web interface is retained as a legacy resource and is not linked here.
Perturbation prediction and AI modeling
Predict single-cell responses across biological contexts
AI models for predicting how genetic perturbations reshape cell state, identity, and population composition.
pertTF
A context-aware transformer model that predicts single-cell transcriptional responses to unseen genetic perturbations in unseen cellular contexts. pertTF also models perturbation-induced changes in cell identity and population composition and supports transfer learning across biological systems.
Single-cell functional genomics
Model heterogeneous perturbation responses
Methods for connecting genetic perturbations with cell states and variable single-cell responses.
PS
The perturbation-response score quantifies heterogeneous responses in single-cell perturbation assays, including Perturb-seq.
scMAGeCK
Models the effects of gene and regulatory-element perturbations across multiple phenotypes in single-cell CRISPR screens.
Guide and editor modeling
Design and evaluate genome-editing experiments
Predictive models and design resources for Cas13, base editing, and focused CRISPR screens.
DeepCas13
A deep-learning model for Cas13d guide activity and off-target effects. The publication is the stable reference for the method.
CRISPR-FOCUS
A strategy for designing focused CRISPR screening libraries with improved coverage of selected gene sets.
BEEP
A prediction tool for base-editing efficiency across candidate guide and editor combinations.
MAGeCK-NEST
Uses protein-interaction information to improve the design and analysis of CRISPR screens.
Transcriptome methods
Foundational RNA-seq software
Earlier methods for transcript assembly, isoform inference, expression estimation, and mutation calling.
Questions about a method?
See the associated publication for method details and cite the primary paper when using a tool.