New Year Offer End Date: 30th April 2024
Skills you will gain:
About Program:
This three-day workshop introduces participants to single-cell multi-omics integration, spatial transcriptomics, AI-based tissue mapping, and biomarker discovery. Through structured lectures, guided demonstrations, and Google Colab exercises, participants will learn how transcriptomic, epigenomic, protein, and spatial datasets can be analysed to identify disease-associated cellular states, tissue microenvironments, regulatory mechanisms, and predictive biomarkers for precision medicine and therapeutic research.
Aim: The aim of this workshop is to provide participants with practical and conceptual knowledge of single-cell multi-omics integration, spatial transcriptomics, AI-based tissue analysis, and machine-learning-assisted biomarker discovery for biomedical, clinical, pharmaceutical, and translational research.
Program Objectives:
- Understand the principles of single-cell and spatial multi-omics technologies.
- Learn the structure and biological relevance of scRNA-seq, scATAC-seq, surface-protein, and spatial transcriptomics datasets.
- Understand preprocessing, normalization, quality control, and batch harmonization.
- Integrate multiple single-cell omics layers using AI-based latent-representation models.
- Identify disease-associated cellular states and regulatory programmes.
- Link chromatin accessibility with gene-expression patterns.
- Understand trajectory inference and pseudotime modelling.
- Compare sequencing-based and imaging-based spatial transcriptomics platforms.
- Map single-cell populations onto tissue architecture.
- Perform spatial deconvolution and cell-type abundance estimation.
What you will learn?
“`html
📅 Day 1: Single-Cell Multi-Omics Integration and AI Latent Representation
- Focus: Integrating single-cell RNA, chromatin-accessibility, and surface-protein data using multimodal AI frameworks to uncover shared cellular states, regulatory programmes, and disease-associated driver biomarkers.
- Introduction to single-cell multi-omics and the biological significance of combining transcriptomic, epigenomic, and surface-protein information.
- Overview of single-cell RNA sequencing, single-cell ATAC sequencing, and CITE-seq-based surface-protein profiling.
- Preprocessing, normalization, feature selection, batch harmonization, and multimodal quality-control strategies.
- Identification of low-quality cells, doublets, technical noise, and modality-specific biases in clinical single-cell datasets.
- Joint latent-space embedding of multiple omics layers using Multi-Omics Factor Analysis.
- Introduction to variational autoencoders for multimodal data integration using scVI and MultiVI.
- Identification of shared and modality-specific latent factors associated with cell identity, disease state, and biological variation.
- Linking scATAC-seq chromatin-accessibility profiles with gene-expression patterns.
- Inference of cell-type-specific regulatory programmes, transcription-factor activity, and gene-regulatory relationships.
- Unsupervised trajectory inference and pseudotime modelling for cellular differentiation and disease progression.
- Identification of latent disease states, regulatory drivers, altered cellular populations, and potential molecular biomarkers.
🛠️ Hands-on:
- Load prepared single-cell RNA-seq and scATAC-seq datasets from a clinical cohort in Google Colab.
- Examine multimodal quality-control metrics and identify technical variation across cells and samples.
- Preprocess and align transcriptomic and chromatin-accessibility datasets for multimodal integration.
- Build a guided multimodal variational autoencoder pipeline using scVI-tools.
- Generate a shared latent representation and visualize integrated cellular populations using dimensionality-reduction plots.
- Compare healthy and disease-associated cellular states and examine latent factors related to disease progression.
- Identify candidate marker genes, regulatory features, and latent disease-associated factors.
🧰 Tools Covered: Google Colab, Python, Scanpy, scVI-tools, MultiVI, MOFA+, Muon, Pandas, NumPy, and Matplotlib
📅 Day 2: High-Resolution Spatial Mapping and AI Deconvolution
- Focus: Mapping single-cell transcriptomic profiles onto physical tissue architecture and estimating spatial cell-type abundances using deep-learning and Bayesian deconvolution methods.
- Introduction to spatial transcriptomics and the importance of spatial context in biomedical and clinical research.
- Comparison of sequencing-based and imaging-based spatial transcriptomics technologies.
- Overview of high-resolution spatial platforms including 10x Genomics Visium HD, Xenium, and MERFISH.
- Understanding spatial gene-expression matrices, tissue coordinates, image features, and spatial metadata.
- Integration of tissue histology with spatial gene-expression measurements.
- Deep-learning-based cell and nucleus segmentation using Cellpose.
- Extraction of cellular, morphological, molecular, and spatial features from tissue sections.
- Principles of spatial spot deconvolution in multicellular tissue regions.
- Bayesian estimation of spatial cell-type abundance using Cell2location.
- Mapping single-cell reference profiles onto tissue sections using Tangram.
- Identification and visualization of spatially variable genes across heterogeneous tissue regions.
- Construction of spatial-neighbourhood graphs and tissue-domain clustering using Squidpy.
- Identification of tumour, immune, stromal, epithelial, and disease-associated tissue domains.
- Mapping tumour boundaries, invasive fronts, immune-enriched regions, and spatially heterogeneous zones.
🛠️ Hands-on:
- Load a prepared spatial transcriptomics dataset and associated high-resolution tissue histology image in Google Colab.
- Visualize selected gene-expression patterns across the tissue section.
- Examine spatial coordinates, tissue regions, and cell-type reference profiles.
- Deconvolve multicellular spatial transcriptomics spots using a guided Cell2location workflow.
- Estimate and visualize cell-type abundances across tumour, stromal, and immune regions.
- Identify spatially variable genes and compare their distribution across tissue domains.
- Generate spatial-domain maps and identify tumour-boundary and immune-enriched zones.
- Interpret changes in cellular composition across disease-associated tissue regions.
🧰 Tools Covered: Google Colab, Python, Squidpy, Cell2location, Tangram, Cellpose, Scanpy, Pandas, NumPy, and Matplotlib
📅 Day 3: Spatial Niche Modelling and AI-Driven Biomarker Discovery
- Focus: Modelling localized cell–cell communication networks, characterizing spatial microenvironments, and identifying predictive biomarkers for patient stratification and clinical translation.
- Introduction to spatial cellular niches, cellular neighbourhoods, and disease-associated tissue microenvironments.
- Biological significance of tumour, immune, stromal, and treatment-response-associated spatial niches.
- Spatial ligand–receptor interaction modelling and localized cell–cell communication analysis.
- Identification of signalling relationships between tumour cells, immune populations, fibroblasts, and stromal cells.
- Cellular-neighbourhood analysis using spatial proximity, cell identity, and gene-expression information.
- Introduction to LIANA+ for analysing and comparing ligand–receptor communication networks.
- Representation of tissue architecture as a graph containing cellular nodes, spatial edges, and molecular features.
- Fundamentals of Graph Neural Networks, including nodes, edges, feature propagation, and neighbourhood aggregation.
- Application of GNNs for modelling tissue microenvironments and cell–cell crosstalk.
- Identification of disease-associated, immune-suppressive, and treatment-response spatial niches.
- Feature selection and machine-learning pipelines for biomarker-signature extraction.
- Classification of treatment responders and non-responders using molecular, cellular, and spatial features.
- Evaluation of predictive models using accuracy, precision, recall, confusion matrix, ROC curve, and AUC.
- Identification of diagnostic, prognostic, predictive, and therapeutic-response biomarkers.
- Validation of spatial and single-cell biomarker panels across patient cohorts.
- Patient stratification, therapeutic-target prioritization, and biological interpretation of biomarker signatures.
- Clinical and translational considerations for converting computational signatures into validated biomarker panels.
🛠️ Hands-on:
- Load a prepared spatial-neighbourhood dataset containing cellular, molecular, and treatment-response information.
- Examine spatial ligand–receptor interaction scores and identify important cell–cell communication patterns.
- Construct a graph representation of spatial cellular neighbourhoods.
- Prepare node features, spatial connections, and treatment-response labels for graph-based modelling.
- Train a guided Graph Neural Network model to classify responder and non-responder tissue environments.
- Select important genes, cell types, ligand–receptor interactions, and spatial features.
- Evaluate model performance using accuracy, confusion matrix, ROC curve, and AUC score.
- Rank important molecular and spatial features to generate a preliminary predictive biomarker signature.
- Interpret the biomarker signature for treatment-response prediction, patient stratification, and therapeutic targeting.
🧰 Tools Covered: Google Colab, Python, Squidpy, LIANA+, PyTorch Geometric, Scikit-learn, Pandas, NumPy, Matplotlib, and SHAP
Mentor Profile
Fee Plan
INR 1999 /- OR USD 50
Get an e-Certificate of Participation!
Intended For :
- Graduate and postgraduate students in biotechnology, bioinformatics, molecular biology, genetics, biochemistry, life sciences, pharmacy, biomedical sciences, computational biology, data science, medicine, and allied disciplines
- PhD scholars, research fellows, project associates, and research assistants working in genomics, cancer biology, immunology, neuroscience, precision medicine, biomarker discovery, or drug discovery
- Academicians, faculty members, research supervisors, principal investigators, and laboratory professionals
- Bioinformaticians, computational biologists, molecular biologists, geneticists, and biomedical researchers
- Industry professionals from biotechnology, pharmaceuticals, genomics, diagnostics, clinical research, precision medicine, healthcare analytics, and AI-driven life-science companies
- Professionals interested in single-cell analysis, spatial transcriptomics, multi-omics integration, biomarker discovery, and computational biology
,
- Graduate and postgraduate students in biotechnology, bioinformatics, molecular biology, genetics, biochemistry, life sciences, pharmacy, biomedical sciences, computational biology, data science, medicine, and allied disciplines
- PhD scholars, research fellows, project associates, and research assistants working in genomics, cancer biology, immunology, neuroscience, precision medicine, biomarker discovery, or drug discovery
- Academicians, faculty members, research supervisors, principal investigators, and laboratory professionals
- Bioinformaticians, computational biologists, molecular biologists, geneticists, and biomedical researchers
- Industry professionals from biotechnology, pharmaceuticals, genomics, diagnostics, clinical research, precision medicine, healthcare analytics, and AI-driven life-science companies
- Professionals interested in single-cell analysis, spatial transcriptomics, multi-omics integration, biomarker discovery, and computational biology
Career Supporting Skills
Program Outcomes
- Explain the principles of spatial transcriptomics and single-cell multi-omics.
- Differentiate between scRNA-seq, scATAC-seq, protein, and spatial datasets.
- Understand essential preprocessing and multimodal quality-control steps.
- Interpret integrated single-cell datasets and latent cellular representations.
- Identify disease-associated cell populations and molecular programmes.
- Understand how chromatin accessibility influences gene expression.
- Interpret cellular trajectories and disease-progression patterns.
- Compare major spatial transcriptomics technologies.
- Visualize gene-expression patterns within tissue architecture.
- Interpret spatial cell-type abundance estimates.
- Identify spatially variable genes and tissue-specific domains.
- Map tumour, immune, stromal, and disease-associated tissue regions.
- Analyse localized cell–cell communication networks.
- Interpret ligand–receptor interactions within spatial niches.
- Understand the role of Graph Neural Networks in tissue modelling.
- Apply machine-learning concepts to biomarker selection.