scRNA analysis (Seurat)

[chris-kreitzer]

06/15: scRNA-analysis

I started with copying all the 4 libraries from the cluster to the local machine
(following Alison's script; with some modifications)

Excluding mitochondrial genes

Work with NV2_ID's rather than gene_short_names, as rownames(lib1) are the IDs and not the gene_short_names (as in Alison's annotation table)

[chris-kreitzer]

``06/17: updates on annotation transformation

The main goal is to provide unique gene_names in the Seurat object, so that we can somehow make sense of downstream analysis.
We have encountered several obstacles throughout this conversion; however, I now present some suggestions.

• I adapt gene_short_name from NVE models whenever a direct match between NVE and NV2 is possible.
• I leave the NV2 ID (mere number) if there is no match with NVE
• In the case that there are several NV2.'s for one NVE (there are obviously duplicated entries in the data.frame), so I just append a, b, etc. (length of duplication to respective gene_short_name so that all entries are getting unique.
• example:

NVE(...) matches NV2.108 & NV2.109; Since we adopt ELK1-like annotation from the NVE, this entry (ELK1-like) is obviously duplicated.
I added now ELK1-like_a for NV2.108 and ELK1-like_b for NV2.109; meaning that both products are essentially the same but got a unique identifier.