MIT CompBio Lecture 21 - Single-cell genomics (Fall 2019)

MIT Computational Biology: Genomes, Networks, Evolution, Health
compbio.mit.edu/6.047/
Prof. Manolis Kellis
Full playlist with all videos in order is here: • Machine Learning in Ge...
All slides from Fall 2019 are here: stellar.mit.edu/S/course/6/fa...
Outline for this lecture:
1. Single-cell profiling technologies
- Traditional single-cell analyses
- Single-cell RNA-seq
- Dealing with noise in scRNA-seq data
- Multiplexing: reduce batch effects, doublets, cost
- Single-cell epigenomics (scATAC-Seq)
- Single-cell multi-omics (PAIRED-seq, SNARE-seq, sci-CAR)
2. Extracting biological insights from single-cell data
- Clustering similar cells
- Clustering similar genes
- Dimensionality reduction
- Distinguishing different cell types
- Trajectories through cell space
- Dataset completion and missing data imputation
- Multiresolution analysis
- Comparison of multiple methods
3. Single-cell RNA-seq in disease: Focus on Brain Disorders
- Why Brain: Cell type and function diversity
- Initial maps of brain diversity across regions, development, organoids
- Brain variation at the single-cell level in Alzheimer’s disease
- Somatic mosaicism and clonality from scDNA-seq and scRNA-seq
- Deconvolution of bulk data into single-cell profiles vs. phenotype vs. genotype
- Deconvolution of eQTL effects at single-cell level and mediation analysis

[@fangfangyan2991]

Great lecture! Thanks for sharing!

[@kumarrahulbhadani]

At 36:55, can you please clarify on clustering together various cells that are nearest neighbors? On what basis we can cluster those fifty cells to form neighbors? Why did you call virtual amalgamations at 37:03? Also, can you please provide any reference reading on this topic?

[@InquilineKea]

Link to 41:50?