Precise reconstruction of the TME using bulk RNA-seq and a machine learning algorithm trained on artificial transcriptomes

Zaitsev, Aleksandr; Chelushkin, Maksim; Dyikanov, Daniiar; Cheremushkin, Ilya; Shpak, Boris; Nomie, Krystle; Zyrin, Vladimir; Nuzhdina, Ekaterina; Lozinsky, Yaroslav; Zotova, Anastasia; Degryse, Sandrine; Kotlov, Nikita; Baisangurov, Artur; Shatsky, Vladimir; Afenteva, Daria; Кузнецов, А. А.; Paul, Susan Raju; Davies, Diane; Reeves, Patrick M.; Lanuti, Michael; Goldberg, Michael F.; Tazearslan, Cagdas; Chasse, Madison; Wang, Iris; Abdou, Mary; Aslanian, Sharon M; Andrewes, Samuel; Hsieh, James J.; Ramachandran, Akshaya; Lyu, Yang; Galkin, Ilia; Svekolkin, Viktor; Cerchietti, Leandro; Poznansky, Mark C.; Ataullakhanov, Ravshan; Fowler, Nathan; Bagaev, Alexander

doi:10.1016/j.ccell.2022.07.006

Cited by 63 publications

(66 citation statements)

References 95 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, IL7R and CCR7 were highly expressed in naive CD4+ T cells, and GZMK and CD160 were highly expressed in cytotoxic T cells. In particular, we found that ImmCluster identified XCL2 and PRF1 as being highly expressed in cytotoxic T cells and NK cells, respectively, which has also been shown in a recent study ( 38 ) but not in the original study (Figure 3C ). These results suggested that ImmCluster can identify novel genes of interest and generate new hypothesis for investigation.…”

Section: Case Studysupporting

confidence: 80%

ImmCluster: an ensemble resource for immunology cell type clustering and annotations in normal and cancerous tissues

Jiang

Zhou

Sheng

et al. 2022

Nucleic Acids Research

View full text Add to dashboard Cite

Single-cell transcriptome has enabled the transcriptional profiling of thousands of immune cells in complex tissues and cancers. However, subtle transcriptomic differences in immune cell subpopulations and the high dimensionality of transcriptomic data make the clustering and annotation of immune cells challenging. Herein, we introduce ImmCluster (http://bio-bigdata.hrbmu.edu.cn/ImmCluster) for immunology cell type clustering and annotation. We manually curated 346 well-known marker genes from 1163 studies. ImmCluster integrates over 420 000 immune cells from nine healthy tissues and over 648 000 cells from different tumour samples of 17 cancer types to generate stable marker-gene sets and develop context-specific immunology references. In addition, ImmCluster provides cell clustering using seven reference-based and four marker gene-based computational methods, and the ensemble method was developed to provide consistent cell clustering than individual methods. Five major analytic modules were provided for interactively exploring the annotations of immune cells, including clustering and annotating immune cell clusters, gene expression of markers, functional assignment in cancer hallmarks, cell states and immune pathways, cell–cell communications and the corresponding ligand–receptor interactions, as well as online tools. ImmCluster generates diverse plots and tables, enabling users to identify significant associations in immune cell clusters simultaneously. ImmCluster is a valuable resource for analysing cellular heterogeneity in cancer microenvironments.

show abstract

Section: Case Studysupporting

confidence: 80%

ImmCluster: an ensemble resource for immunology cell type clustering and annotations in normal and cancerous tissues

Jiang

Zhou

Sheng

et al. 2022

Nucleic Acids Research

View full text Add to dashboard Cite

show abstract

“…Numerous translational research studies have demonstrated that PD-L1 expression, TMB (with clonal TMB showing increased predictive ability vs. TMB methods including all somatic mutations), and other immune related gene expression markers focusing on the tumor microenvironment (TME) are independent predictors of response 15,[34][35][36][37][38][39][40][41][42][43][44][45][46][47] . For example, in bladder cancer, multiple studies have demonstrated the potential for PD-L1 by IHC, TMB, and T-cell-inflamed gene expression to predict PD-(L)1 therapy benefit, whether alone or in combination with chemotherapy, with an only increasing need to maximize PD-(L)1 benefit given the number of other approved agents in different therapy classes (chemotherapy, antibody drug conjugates and small molecule inhibitors) that must sequenced [48][49][50][51][52][53] .…”

Section: Plain Language Summarymentioning

confidence: 99%

Development and validation of an integrative pan-solid tumor predictor of PD-1/PD-L1 blockade benefit

Tomlins¹,

Khazanov²,

Bulen³

et al. 2023

Commun Med

View full text Add to dashboard Cite

Background Anti-PD-1 and PD-L1 (collectively PD-[L]1) therapies are approved for many advanced solid tumors. Biomarkers beyond PD-L1 immunohistochemistry, microsatellite instability, and tumor mutation burden (TMB) may improve benefit prediction. Methods Using treatment data and genomic and transcriptomic tumor tissue profiling from an observational trial (NCT03061305), we developed Immunotherapy Response Score (IRS), a pan-tumor predictive model of PD-(L)1 benefit. IRS real-world progression free survival (rwPFS) and overall survival (OS) prediction was validated in an independent cohort of trial patients. Results Here, by Cox modeling, we develop IRS—which combines TMB with CD274, PDCD1, ADAM12 and TOP2A quantitative expression—to predict pembrolizumab rwPFS (648 patients; 26 tumor types; IRS-High or -Low groups). In the 248 patient validation cohort (248 patients; 24 tumor types; non-pembrolizumab PD-[L]1 monotherapy treatment), median rwPFS and OS are significantly longer in IRS-High vs. IRS-Low patients (rwPFS adjusted hazard ratio [aHR] 0.52, p = 0.003; OS aHR 0.49, p = 0.005); TMB alone does not significantly predict PD-(L)1 rwPFS nor OS. In 146 patients treated with systemic therapy prior to pembrolizumab monotherapy, pembrolizumab rwPFS is only significantly longer than immediately preceding therapy rwPFS in IRS-High patients (interaction test p = 0.001). In propensity matched lung cancer patients treated with first-line pembrolizumab monotherapy or pembrolizumab+chemotherapy, monotherapy rwPFS is significantly shorter in IRS-Low patients, but is not significantly different in IRS-High patients. Across 24,463 molecularly-evaluable trial patients, 7.6% of patients outside of monotherapy PD-(L)1 approved tumor types are IRS-High/TMB-Low. Conclusions The validated, predictive, pan-tumor IRS model can expand PD-(L)1 monotherapy benefit outside currently approved indications.

show abstract

“…Further assessment of cellular heterogeneity used two additional methodologies that can differentiate the cell types present in a tumor: deconvolution of bulk RNA-seq using a custom algorithm (Zaitsev et al, 2022) (Figure 4C; Table S3) and (F) Boxplots showing the relative cell-type percentage for tumor lesions and margins (10 samples from four patients); endothelial cell content is presented as a relative area of the CD31 marker mask. In the boxplots, the right whisker indicates the maximum value or 75th percentile + 1.5 interquartile range (IQR); the left whisker indicates the minimum value or 25th percentile À 1.5 IQR.…”

Section: Cytof Analysis Revealed Striking Similarity In Immune Cell C...mentioning

confidence: 99%

“…The recently-described machine learning algorithm, Kassandra, was used to predict cell percentages from bulk RNA-seq (Zaitsev et al, 2022). The model consisted of a two-level hierarchical ensemble that used LightGBM as building blocks.…”

Section: Deconvolution Of Bulk Rna-seqmentioning

confidence: 99%

Precise reconstruction of the TME using bulk RNA-seq and a machine learning algorithm trained on artificial transcriptomes

Cited by 63 publications

References 95 publications

ImmCluster: an ensemble resource for immunology cell type clustering and annotations in normal and cancerous tissues

ImmCluster: an ensemble resource for immunology cell type clustering and annotations in normal and cancerous tissues

Development and validation of an integrative pan-solid tumor predictor of PD-1/PD-L1 blockade benefit

Multiregional single-cell proteogenomic analysis of ccRCC reveals cytokine drivers of intratumor spatial heterogeneity

Contact Info

Product

Resources

About