SC3 - consensus clustering of single-cell RNA-Seq data

A talk that was given at the internal Genome Campus seminar at the European Bioinformatics Institute.

Transcript

1. SC3 - consensus clustering of single-cell RNA-Seq data Vladimir Kiselev PostDoc Hemberg group
2. Identification of cell types ~45000 cells from mouse retina Wikipedia: “Animals have evolved a greater diversity of cell types in a multicellular body (100–150 different cell types)” Old methods • Surface markers • Morphology New methods • Single-cell RNA-Seq
3. Unsupervised clustering of cells Facts: • More than 100 clustering algorithms available • Single-Cell data is new and high-dimensional • Standard robust and efficient algorithm is k-means Problems with new algorithms: • Parameters • Speed • Scalability
4. Distance Dimensionality reduction PCA Spectral MDS Spectral Reg. Pearson Spearman Euclidean Minkowski Manhattan Gene Filter Genes Cell Filter d - first d eigenvectors N Cells reduction of dimensionality k-means k clusters k is known! d Dimensionality reduction pipeline
5. define k; k-means
6. define k; (randomly generate k centroids) k-means
7. define k; (randomly generate k centroids) cluster by nearest centroid readjust centroids k-means
8. define k; for(i = 0; i < number of starts; i++) { (randomly generate k centroids) for(j = 0; j < number of iterations; j++) { cluster by nearest centroid readjust centroids } } } k-means
9. Adjusted Rand Index (ARI) k-means ARI • Like Spearman correlation between two clusterings • If ARI = 0.8 then clustering is very good d Distance Dimensionality reduction Pearson Spearman Euclidean Minkowski Manhattan Gene Filter Genes Cell Filter d - first d eigenvectors N Cells reduction of dimensionality k clusters gold standard is known! PCA Spectral MDS Spectral Reg.
10. Datasets for pipeline testing Publication N k Name Treutlein, B. et al. Reconstructing lineage hierarchies of the distal lung epithelium using single-cell RNA-seq. Nature 509, 371–375 (2014) 80 5 quake Ting, D. T. et al. Single-cell RNA sequencing identifies extracellular matrix gene expression by pancreatic circulating tumor cells. Cell Rep. 8, 1905–1918 (2014). 149 7 ting Deng, Q., Ramsköld, D., Reinius, B. & Sandberg, R. Single-cell RNA-seq reveals dynamic, random monoallelic gene expression in mammalian cells. Science 343, 193–196 (2014) 268 10 sandberg Pollen, A. A. et al. Low-coverage single-cell mRNA sequencing reveals cellular heterogeneity and activated signaling pathways in developing cerebral cortex. Nat. Biotechnol. 32, 1053–1058 (2014). 301 11 pollen Patel, A. P. et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science 344, 1396–1401 (2014) 430 5 bernstein Usoskin, D. et al. Unbiased classification of sensory neuron types by large-scale single- cell RNA sequencing. Nat. Neurosci. 18, 145–153 (2015). 622 11 usoskin Klein, A. M. et al. Droplet Barcoding for Single-Cell Transcriptomics Applied to Embryonic Stem Cells. Cell 161, 1187–1201 (2015). 2717 4 kirschner Zeisel, A. et al. Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science 347, 1138–1142 (2015) 3005 9 linnarsson Macosko, E. Z. et al. Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell 161, 1202–1214 (2015). 44808 39 maccarroll