No abstract

Chromatin accessibility is essential for transcriptional activation of genomic regions. It is well established that transcription factors (TFs) and histone modifications (HMs) play critical roles in chromatin accessibility regulation. However, there is a lack of studies that quantify these relationships. Here we constructed a two-layer model to predict chromatin accessibility by integrating DNA sequence, TF binding, and HM signals. By applying the model to two human cell lines (GM12878 and HepG2), we found that DNA sequences had limited power for accessibility prediction, while both TF binding and HM signals predicted chromatin accessibility with high accuracy. According to the HM model, HM features determined chromatin accessibility in a cell line shared manner, with the prediction power attributing to five core HM types. Results from the TF model indicated that chromatin accessibility was determined by a subset of informative TFs including both cell line-specific and generic TFs. The combined model of both TF and HM signals did not further improve the prediction accuracy, indicating that they provide redundant information in terms of chromatin accessibility prediction. The TFs and HM models can also distinguish the chromatin accessibility of proximal versus distal transcription start sites with high accuracy.

Introduction: Accumulating evidence has suggested that the tumor immune microenvironment (TIME) drastically impacts cancer patients’ clinical outcomes, especially immunotherapy response. Many transcriptome-based cell-type quantification methods (e.g., TIMER, CIBERSORTx) have been developed to estimate immune cell infiltration from bulk RNA sequencing. Our aim is to reveal the cell-cell interaction and its association with clinical outcomes. Methods: We developed, TimiGP (Tumor Immune Microenvironment Illustration based on Gene Pairing), a computational method to investigate the TIME by inferring the inter-cell “functional” interaction network and estimating the clinical value of immune cells from bulk gene expression and clinical information. It was based on two assumptions: 1) The TIME is a dynamic balance between pro-and anti-tumor immune cells, whose relative functions determine the capacity of the immune system and subsequently impact clinical outcomes; 2) The pairwise relation between gene pairs can capture the relative functions between immune cells. TimiGP was applied to study the association of TIME with prognosis (7,938 samples, 23 cancer types) and the immunotherapy response (2,941 patients, 5 cancer types, 6 treatments) in pan-cancer analysis. Results: We first applied 8 existing deconvolution methods to highly immune infiltrated metastatic melanoma (MM), whose results identified both anti- and pro-tumor cell types (e.g., tumor-associated neutrophils) associated with a superior prognosis. By contrast, TimiGP overcame such prognostic bias and identified the prognostic value of immune cells consistent with their established functions. We then integrated TimiGP with single-cell RNA-seq analysis to achieve higher resolution for specific immune cell subsets than existing methods. Furthermore, we utilized TimiGP in pan-cancer analysis, which identified the TIME heterogeneity associated with survival and critical immune interactions determining the immunotherapy response across various cancer types and treatments. Besides, TimiGP could integrate the TIME insights to construct computational-friendly and interpretable models for risk evaluation or response prediction. For example, we used it to develop MM prognostic model that achieved superior performance in 5 independent validations (C-index: 0.621-0.659) than traditional methods. Conclusions: TimiGP (R package available at https://github.com/CSkylarL/TimiGP) is a novel algorithm that accepts bulk expression and clinical information to generate a prognostic/responsive score of each cell type and an inter-cell functional interaction network. It can be used to identify target-to-hit, study the underlying causes of clinical heterogeneity, and construct predictive models based on biological and clinical insights, which offer substantial potential for personalized cancer treatment. Citation Format: Chenyang Li, Wei Hong, Baoyi Zhang, Evelien Schaafsma, Alexandre Reuben, Jianjun Zhang, Chao Cheng. TimiGP: Dissect the tumor immune microenvironment and its association with survival and immunotherapy response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2080.