Trascriptomics and AI approaches for genome analysis

Explore the cutting-edge intersection of transcriptomics and artificial intelligence. Learn to predict cellular responses to perturbations, analyze single-cell RNA sequencing data, and build virtual cell models that simulate biological behavior without traditional wet-lab experiments.

Credits

3 Credits (Lecture)

Level

300 + 700

Format

Online - synchronous
Tuesdays 2:30-5:00pm

Class Size

Limited to 35 students

Prerequisites: BIO 203

Weeks 1-2: Transcriptomics Fundamentals Central dogma review, gene expression and regulation, RNA biogenesis, alternative splicing, RNA-seq technologies (bulk and single-cell), library preparation methods, sequencing platforms, and transcriptome data structure

Weeks 3-4: Single-Cell Data Analysis Quality control, normalization, dimensionality reduction, clustering, and differential expression analysis

Weeks 5-6: Perturbation Approaches in Genomics Gene perturbation strategies, Perturb-seq methods, linking genotype to phenotype, high-throughput functional screens, and interpreting perturbation data

Weeks 7-9: AI for Biology Machine learning fundamentals, deep learning architectures, transformer models, protein language models (ESM2), biological embeddings, attention mechanisms, and introduction to the Arc Institute’s STATE model

Weeks 10-11: Virtual Cell Modeling Predicting perturbation responses, context generalization across cell types, causal vs. correlational modeling

Week 12: Large-Scale Datasets & Resources Arc Virtual Cell Atlas, CELL×GENE, scBaseCount, Tahoe-100M dataset exploration

Weeks 13-14: Model Evaluation & Benchmarking Perturbation discrimination, differential expression metrics, MAE, and Virtual Cell Challenge framework

Week 15: Final Exam [.description]##

Students will be able to:

Analyze scRNA-seq Data Build Predictive Models Design Perturbation Experiments Interpret AI Predictions Use Large Scale Datasets Apply Virtual Cell Models

Learn to work with real transcriptomic datasets including single-cell RNA-seq profiles. Master quality control pipelines, normalization techniques, and batch effect correction. Perform dimensionality reduction (PCA, UMAP, t-SNE), cell type identification, and trajectory analysis. Apply differential expression analysis across conditions and cell types. Explore virtual cell models using the Arc Institute’s STATE architecture and the Virtual Cell Challenge framework.

Instructor: Assoc. Prof. Konstantinos Krampis
Contact: kk104@hunter.cuny.edu