MSofTimiGP-Response

This repo includes all codes and intermediate for the manuscript of the TimiGP-Response. Of note, it includes the cell-cell interaction network of each dataset for the pan-cancer immune landscape associated with immunotherapy response.

Table of Contents

• TimiGP-Response
• Overview of the Repository
• Dataset Description
• Pan-Cancer Immune Landscape
• Citation
• LICENSE

TimiGP-Response

TimiGP-Response is an analysis module under the TimiGP computaional framework. It is designed to infer cell-cell interactions in tumor immune microenvironment (TIME) through gene pairs, and evaluate the cell types assocaited with immunotherapy response.

For more details, please read our manuscript: TimiGP-Response: the pan-cancer immune landscape associated with response to immunotherapy..

Overview of the Repository

This repository contains the codes and intermediate data for the manuscript of the TimiGP-Response. The repository is organized as follows:

1. data folder: The data used in this study are publicly available. Please follow the instruction to download datasets to complete the data folder.

2. Fig1 folder: The codes and intermediate data for Fig1.

   1. code folder: config files and R codes to generate the corresponding figures.

      • Fig1_config.R file: config file for the R codes.
      • Fig1_function.R file: function file for the R codes.
      • Fig1.*.R files: R codes to generate the corresponding figures. Please run following the order of the file name (from script 1 to 5).
      • sessionInfo.txt file: the session information of the R environment.

   2. result folder: intermediate results and original figures.

      • TNBC_scRNA_GSE169246 folder: results generated from Fig1.script1_Preprocess_scRNAseq.R
      • customized_markers_for_TimiGP folder: results generated from Fig1.script2_generate_TimiGP_markers.R
      • TimiGP_TNBC_scRNA_marker_for_spatial folder: results generated from Fig1.script3_customized_TimiGP_response_for_TNBC.R
      • TNBC_spatial_ratio_responder folder: results generated from Fig1.script4_reanalyze_spatial_IMC.R
      • Figures_Tables folder: results generated from Fig1.script5_generate_Figures_Tables.R

3. Fig2 folder: The codes and intermediate data for Fig2.

   1. code folder: config files and R codes to generate the corresponding figures.
      • config.rda file: config file for the R codes.
      • Fig2.*.R files: R codes to generate the corresponding figures. Please run following the order of the file name (from script 1 to 5).
      • sessionInfo.txt file: the session information of the R environment.
   2. result folder: intermediate results and original figures.
      • pan_cancer_Tumor_TIME_control folder: results generated from Fig2.script1_pan_cancer_Tumor_TIME_control.R
      • pan_cancer_TIME_landscape folder: results generated from Fig2.script2_pan_cancer_TIME_landscape.R
      • pan_cancer_T_Cell_landscape folder: results generated from Fig2.script3_pan_cancer_T_Cell_landscape.R
      • summary folder: results generated from Fig2.script4_summary_favorability_score.R and Fig2.script5_summary_TIME_similarity.R

Dataset Description

Here is a summary of the pan-cancer immunotherapy datasets (bulk transcriptomics) used in this study:

Cancer Type	Datasets	Targets	No. Samples	No. Non-responder (0)	No. Responder (1)	Drug	Therapy Type	Reference
SKCM	phs000452.v2.p1: IPI	CTLA4	37	23	14	Ipilimumab	Immunotherapy	Van Allen et al.
SKCM	GSE35640: MAGE-3	MAGE-A3	56	34	22	MAGE-A3 Vaccine	Immunotherapy	Ulloa-Montoya et al.
SKCM	GSE78220: NIV/PEM	PD-1	27	12	15	Nivolumab/Pembrolizumab	Immunotherapy	Hugo et al.
SKCM	phs000452.v3.p1: NIV/PEM	PD-1	121	72	49	Nivolumab/Pembrolizumab	Immunotherapy	Liu et al.
SKCM	Gide2019: NIV/PEM+(IPI)	PD-1+(CTLA4)	91	42	49	Nivolumab/Pembrolizumab+ (Ipilimumab)	Immunotherapy	Gide et al.
SKCM	Gide2019: NIV/PEM	PD-1	50	27	23	Nivolumab/Pembrolizumab	Immunotherapy	Gide et al.
SKCM	Gide2019: NIV/PEM+IPI	PD-1+CTLA4	41	15	26	Nivolumab/Pembrolizumab+Ipilimumab	Immunotherapy	Gide et al.
SKCM	GSE91061pre: NIV/PEM+(IPI)	PD1+(CTLA4)	34	25	9	Nivolumab/Pembrolizumab+(Ipilimumab)	Immunotherapy	Riaz et al.
SKCM	GSE91061pre: NIV/PEM	PD-1	15	10	5	Nivolumab/Pembrolizumab	Immunotherapy	Riaz et al.
SKCM	GSE91061pre: NIV/PEM+IPI	PD-1+CTLA4	19	15	4	Nivolumab/Pembrolizumab+Ipilimumab	Immunotherapy	Riaz et al.
SKCM	GSE91061on: NIV/PEM+(IPI)	PD1+(CTLA4)	40	30	10	Nivolumab/Pembrolizumab+(Ipilimumab)	Immunotherapy	Riaz et al.
NSCLC	POPLAR: TXT	NA	75	64	11	Docetaxel	Chemotherapy	Banchereau et al.
NSCLC	OAK: TXT	NA	344	302	42	Docetaxel	Chemotherapy	Patil et al.
NSCLC	GSE166449: PEM	PD-1	22	15	7	Pembrolizumab	Immunotherapy	Lee et al.
NSCLC	GSE126044: NIV/PEM	PD-1	16	11	5	Nivolumab/Pembrolizumab	Immunotherapy	Cho et al.
NSCLC	PCD4989g: ATEZO	PD-L1	54	38	16	Atezolizumab	Immunotherapy	Banchereau et al.
NSCLC	POPLAR: ATEZO	PD-L1	81	70	11	Atezolizumab	Immunotherapy	Banchereau et al.
NSCLC	OAK: ATEZO	PD-L1	339	291	48	Atezolizumab	Immunotherapy	Patil et al.
RCC	IMmotion150: SUN	RTKs	85	57	28	Sunitinib	Targeted Therapy	Banchereau et al.
RCC	IMmotion151: SUN	RTKs	378	239	139	Sunitinib	Targeted Therapy	Motzer et al.
RCC	phs001493.v1.p1:NIV	PD-1	24	19	5	Nivolumab	Immunotherapy	Miao et al.
RCC	IMmotion150: ATEZO	PD-L1	77	62	15	Atezolizumab	Immunotherapy	Banchereau et al.
RCC	PCD4989g: ATEZO	PD-L1	58	50	8	Atezolizumab	Immunotherapy	Banchereau et al.
RCC	IMmotion150: ATEZO+BEV	PD-L1+VEGF	85	54	31	Atezolizumab + Bevacizumab	Immunotherapy + Targeted Therapy	Banchereau et al.
RCC	IMmotion151: ATEZO+BEV	PD-L1+VEGF	380	230	150	Atezolizumab + Bevacizumab	Immunotherapy + Targeted Therapy	Motzer et al.
mUC	IMvigor210: ATEZO	PD-L1	208	163	45	Atezolizumab	Immunotherapy	Banchereau et al.
mUC	PCD4989g: ATEZO	PD-L1	94	72	22	Atezolizumab	Immunotherapy	Banchereau et al.
mUC	Snyder2017: ATEZO	PD-L1	21	14	7	Atezolizumab	Immunotherapy	Snyder et al.
HCC	GO30140+IMbrave150: SOR	RTKs	40	30	10	Sorafenib	Targeted Therapy	Zhu et al.
HCC	GO30140+IMbrave150: ATEZO	PD-L1	49	38	11	Atezolizumab	Immunotherapy	Zhu et al.
HCC	GO30140+IMbrave150: ATEZO+BEV	PD-L1+VEGF	255	169	86	Atezolizumab + Bevacizumab	Immunotherapy + Targeted Therapy	Zhu et al.
BC	GSE194040: PTX (TNBC)	NA	61	43	18	Paclitaxel	Chemotherapy	Wolf et al.
BC	GSE194040: PTX+PEM (TNBC)	PD-1	26	13	13	Paclitaxel + Pembrolizumab	Immunotherapy + Chemotherapy	Wolf et al.
BC	GSE194040: PTX (non-TNBC)	NA	118	81	37	Paclitaxel	Chemotherapy	Wolf et al.
BC	GSE194040: PTX+PEM (non-TNBC)	PD-1	43	32	11	Paclitaxel + Pembrolizumab	Immunotherapy + Chemotherapy	Wolf et al.
EC	GSE165252pre:ATEZO	PD-L1	32	20	12	Atezolizumab	Immunotherapy + Chemoradiotherapy	Van Den Ende et al.
EC	GSE165252on:ATEZO	PD-L1	29	20	9	Atezolizumab	Immunotherapy + Chemoradiotherapy	Van Den Ende et al.
EC	GSE165252post:ATEZO	PD-L1	10	8	2	Atezolizumab	Immunotherapy + Chemoradiotherapy	Van Den Ende et al.

Pan-Cancer Immune Landscape

The pan-cancer immune landscape associated with immunotherapy response is generated from the bulk transcriptomics data using the TimiGP-Response module. They are analyzed at different resolutions as below:

• Pan-Cancer TIME and Tumor Control
• Pan-Cancer TIME Landscape
• Pan-Cancer T Cell Landscape
• Detailed Result per Dataset

Pan-Cancer TIME and Tumor Control

We utilized cell-type markers adapted from Bindea et al., encompassing two critical cell types: cytotoxic cells, serving as positive controls indicative of anti-tumor activity, and tumor cells, designated as negative controls associated with non-responders.

In below figure, the circle plot (left) is a cell-cell interaction example and the scatter pie chart (right) displays the favorability score exported by TimiGP-Response, which estimates the association of Immune cells and controls (x-axis) with immunotherapy responders (favorable score) and non-responders (unfavorable score).

As a result, T cells and cytotoxic cells demonstrated a consistent association with immunotherapy responders, while tumor cells predominated in association with non-responders across nearly all datasets. These results align with the designed controls and the rationale behind immunotherapy targeting T cell responses.

For the cell-cell interaction network and favorability score of each dataset, please go to Detailed Result per Dataset and click the corresponding dataset at this resolution.

Pan-Cancer TIME Landscape

Given the potential bias arising from assigning cytotoxic markers of CD8 T cells and NK cells to the specific cytotoxic cell, we subsequently portraied the TIME utilizing the modified LM22 signature. This signature includes activating and resting immune cell states and has undergone extensive validation.

In below figure, the circle plot (left) is a cell-cell interaction example and the scatter pie chart (right) displays the favorability score exported by TimiGP-Response, which estimates the association of Immune cells (x-axis) with immunotherapy responders (favorable score) and non-responders (unfavorable score).

As a result, the major immunotherapy target, CD8 T cells, along with resting and activated CD4 memory T cells, are consistently associated with responders. Conversely, anti-inflammatory (M2) macrophages and resting mast cells are associated with non-responders to immunotherapy.

For the cell-cell interaction network and favorability score of each dataset, please go to Detailed Result per Dataset and click the corresponding dataset at this resolution.

Pan-Cancer T Cell Landscape

Given that the main target of the immunotherapy in these datasets is T cells, and our analysis at the TIME resolution also highlights the importance of T cells in treatment response, we next focus on T cells for a higher resolution, including 40 T cell subtypes as defined in a pan-cancer T cell scRNA-seq study.

In below figure, the circle plot (left) is a cell-cell interaction example and the scatter pie chart (right) displays the favorability score exported by TimiGP-Response, which estimates the association of T cell subtypes (x-axis) with immunotherapy responders (favorable score) and non-responders (unfavorable score).

As a result, CD8+GZMK+ exhausted T cells (Tex) and CD8+terminal Tex emerged as pivotal cell types associated with responders across nearly all cancer types and immunotherapies in the analysis. CD8+ and CD4+ GZMK+ effector memory T cells (Tem) were also identified as associated with immunotherapy responders. In addition, CD4+IFNG+ follicular/type 1 dual helper T cells (Tfh/Th1) and CD4+TNF+ T cells were identified to demonstrate a favorable correlation with immunotherapy response. As for immunotherapy non-responders, CD8+Tc17 (IL-17 producing CD8+ T cells) emerges as the top candidate.

For the cell-cell interaction network and favorability score of each dataset, please go to Detailed Result per Dataset and click the corresponding dataset at this resolution.

Detailed Result per Dataset

The cell-cell interaction network associated with immunotherpay responders and the favorability score to prioritize the cell types associated with the responders and non-responders for each dataset can be found in the Fig2 folder and accessed through the following links:

Cancer Type	Datasets	TIME and Tumor Control	TIME landscape	T Cell landscape
SKCM	phs000452.v2.p1: IPI	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE35640: MAGE-3	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE78220: NIV/PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	phs000452.v3.p1: NIV/PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	Gide2019: NIV/PEM+(IPI)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	Gide2019: NIV/PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	Gide2019: NIV/PEM+IPI	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE91061pre: NIV/PEM+(IPI)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE91061pre: NIV/PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE91061pre: NIV/PEM+IPI	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
SKCM	GSE91061on: NIV/PEM+(IPI)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	POPLAR: TXT	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	OAK: TXT	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	GSE166449: PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	GSE126044: NIV/PEM	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	PCD4989g: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	POPLAR: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
NSCLC	OAK: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	IMmotion150: SUN	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	IMmotion151: SUN	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	phs001493.v1.p1:NIV	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	IMmotion150: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	PCD4989g: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	IMmotion150: ATEZO+BEV	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
RCC	IMmotion151: ATEZO+BEV	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
mUC	IMvigor210: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
mUC	PCD4989g: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
mUC	Snyder2017: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
HCC	GO30140+IMbrave150: SOR	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
HCC	GO30140+IMbrave150: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
HCC	GO30140+IMbrave150: ATEZO+BEV	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
BC	GSE194040: PTX (TNBC)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
BC	GSE194040: PTX+PEM (TNBC)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
BC	GSE194040: PTX (non-TNBC)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
BC	GSE194040: PTX+PEM (non-TNBC)	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
EC	GSE165252pre: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
EC	GSE165252on: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score
EC	GSE165252post: ATEZO	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score	Cell-Cell Interaction Network & Favorability Score

Citation

This repo is intended for research use only.

If you use TimiGP-Response in your publication, please cite the paper: Li, C. et al. TimiGP-Response: the pan-cancer immune landscape associated with response to immunotherapy. bioRxiv, 2024.2006.2021.600089, doi:10.1101/2024.06.21.600089 (2024).

LICENSE

This repository is licensed under the GPL-3 License. See LICENSE for more information.