Single Cell RNAseq Analysis of Tabula Muris Data

I have made all the notebooks using Kaggle kernel and they can also be viewed there.

Dataset link - https://www.kaggle.com/aayush9753/singlecell-rnaseq-data-from-mouse-brain

• Tabula Muris is a compendium of single-cell transcriptome data from the model organism Mus musculus, containing nearly 100,000 cells from 20 organs and tissues.
• The data allow for direct and controlled comparison of gene expression in cell types shared between tissues, such as immune cells from distinct anatomical locations.
• This data consists of an expression matrix where each column corresponds to a gene (or transcript) and each row corresponds to a single cell and a table of metadata describing each cell.

0. Theory - Introduction to single-cell RNA-seq - Kaggle Link

• Covers most of the theoretical concepts about Bulk RNA-seq, scRNA-seq and Computational Analysis.

1. AnnData and Preprocessing spike-ins - Kaggle Link

• Theory behind AnnData objects
• Exploring a test anndata
• Creating Anndata from the csv files in the dataset
• Preprocessing the labeling spike-ins

2. Quality Control in Single cell RNA-seq data - Kaggle Link

• Calculated the quality control metrices across cells and genes.
• Quality control in cells by removing cells with less total gene count, less unique genes and giving more spike-ins.
• Quality control in genes by removing genes which occur in less unique cells as well as those genes having less total cells.

3. Normalization & PCA - Kaggle Link

• Directly applying PCA on quality controlled data
• Applying CPM normalization and then PCA
• Normalizing each cell by total counts over all genes and excluding highly_expressed genes and then PCA
• Removing offending gene Rn45s and then PCA
• Final, Normalization by log1p and scaling and then PCA.

4. Dimensionality reduction and Clustering - Kaggle Link

1. Dimensionality Reduction

• tSNE
• UMAP

2. Clustering

• k-Means on tSNE ** Evaluating the k-means clustering ** Playing with No of cluster in k-means
• Graph Based Clustering Method - Louvain ** Tuning thr resolution parameter
• Seeing clusters in cells of a perticular sybtissue

5. Differential expression - Kaggle Link

• Building intuition of differential expression
• Ttest
• Working with whole dataset using original labels - "cell_ontology_class"
• Working with whole dataset using "louvain" clusters
• Comapring to known marker genes of call classes