Analysis results

Quality control

## An object of class Seurat 
## 37352 features across 197 samples within 2 assays 
## Active assay: RNA (18676 features, 0 variable features)
##  8 layers present: counts.LSCC1, counts.LSCC2, counts.RCC1, counts.RCC2, data.LSCC1, data.LSCC2, data.RCC1, data.RCC2
##  1 other assay present: originalexp

Preprocessing and Vizualization

Quality control

CCAIntegration

Gene signatures

Here, we selected gene signatures from Azimuth refrence website - https://azimuth.hubmapconsortium.org/references/

Marker genes (TissueSub)

Unsupervised clustering

## Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck
## 
## Number of nodes: 197
## Number of edges: 7464
## 
## Running Louvain algorithm...
## Maximum modularity in 10 random starts: 0.4192
## Number of communities: 4
## Elapsed time: 0 seconds

Marker genes (unsupervised)

Pseudotime (Basic, UMAP)

Here, we conduct a basic pseudotime analysis to explore how effectively the UMAP principal components delineate cells along a differentiation trajectory. To achieve this, we concentrate solely on the TLS and LN spots, recalculating the variable genes within this specific group, and subsequently performing PCA and UMAP. Note: the direction of the trajectory is relative.

Pseudotime (Slingshot, unsupervised)

Here, we conduct a more advanced pseudotime analysis using Slingshot to explore how effectively cells delineate along the sling pseudotime trajectory. For this we build upon the previous basic analysis. Note: the direction of the trajectory is relative. We used Slingshot in unsupervised setting.

Gene Pseudotime (unsupervised)

Here we take the 1000 most variable genes within the TLS and LN group and correlate them with the previously precomputed pseudotime using Slingshot. Then we select the top 50 by fitting a GAM model for each gene using pseudotime as independent variable.

## SSR4 PKN3 IGKC AGPAT4 XBP1NULL

## IGHG2 IGHG4 IGHG3 IGHG1 SCN9ANULL

## IGHA1 IGLL5 DERL3 MZB1 JCHAINNULL

## PIM2 HMGA1 HMGN2 OXT CR2NULL

## H3C15 H1-5 H2AC11 H4C8 FCER2NULL

## H2BC18 H3C13 VPREB3 H3C2 H3C10NULL

## EZR H4C15 DTX1 HMGB2 H3C3NULL

## H2AC13 H2AC21 CD22 H4C6 H3C7NULL

## NIBAN3 H2AC12 P2RX5 H4C4 MARCKSL1NULL

## H3C4 H4C12 H2AZ1 H2AC14 FDCSPNULL

Pseudotime (Slingshot, supervised)

Here, we conduct a more advanced pseudotime analysis using Slingshot to explore how effectively cells delineate along the sling pseudotime trajectory. For this we build upon the previous basic analysis. Note: the direction of the trajectory is relative. We used Slingshot in supervised setting.

Gene Pseudotime (supervised)

Here we take the 1000 most variable genes within the TLS and LN group and correlate them with the previously precomputed pseudotime using Slingshot. Then we select the top 50 by fitting a GAM model for each gene using pseudotime as independent variable.

## PFN1 H4C8 CFL1 CLU CR2NULL

## OAZ1 HMGA1 CORO1A H1-2 TUBA1CNULL

## FDCSP H4C12 H2BS1 MS4A1 SPIBNULL

## H2AC11 H1-4 H4C3 H3C10 H3C2NULL

## GGA2 H3C15 CD37 H4C15 SYNGR2NULL

## LIMD2 IRF8 MCM5 H4C4 H3C13NULL

## H1-5 H2BC5 CD79A H2AC19 ATP2A3NULL

## PRDX1 HLA-DMA NACA NCF1 RAC2NULL

## SERF2 GRB2 CD22 EZR H4C2NULL

## SPARC H2BC8 H1-3 ARPC4 POU2F2NULL