Dissection of 4E and 4D, 7H and 8H

[beyondpie]

During the integration with snm3C, we noticed that the 4E and 4D dissections, 7H and 8H dissections are not consistent with snm3C ones.

After discussion, we list the principles to distinguish the dissections.

4E vs 4D:

1. 4D could have more OGC cells since it’s closed to the white part
2. 4E is closed to 4H, and would have some cells belong to PAL (4H) since the boundaries between 4E and 4H are not that easy to distinguish; but 4D is far away from 4H.
3. Anterior of 4E is 3E

7H vs 8H:

1. anterior to posterior: 4F (SI and MA) - 7H - 8H - 9G
2. MEA region is enriched in 8H (compared with 7H)

From our observations based on the cellMeta (2024-09-14):

1. Our 4E dissection should be 4D and 4D should 4E.

• 4E have more OGC than 4D (1368, 1117) from our data
• 4D is closed to 4H. (cell type proportion spearman correlation: 0.68 between 4D and 4H, 0.29 between 4E and 4H)
• 4D is closer to 3E; and 4D(CP-1) is also closer to ACB region than 4E (ACB-2 ideally).

So 4D should be 4E in our sample.

2. 7H and 8H are OK, no need to change.

• 4F-7H-8H-9G follow the anterior to posterior pattern
  https://www.nature.com/articles/s41586-023-06824-9/figures/8
• 8H has more cells labels having MEA regions than 7H.
  (4823 in 8H, 2977 in 7H, 4035 in 9G).

[beyondpie]

Here I only check the 4D and 4E by mixing the two replicates samples in each of them. We should further check each of the replicate to see if we need to change all the 4D/4E or only one sample.