Immunoinformatics (Amsterdam, Netherlands)最新文献

Response to cancer immunotherapies depends on the complex and dynamic interactions between T cell recognition and killing of cancer cells that are counteracted through immunosuppressive pathways in the tumor microenvironment. Therefore, while measurements such as tumor mutational burden provide biomarkers to select patients for immunotherapy, they neither universally predict patient response nor implicate the mechanisms that underlie immunotherapy resistance. Recent advances in single-cell RNA sequencing technology measure cellular heterogeneity within cells of an individual tumor but have yet to realize the promise of predictive oncology. In addition to data, mechanistic multiscale computational models are developed to predict treatment response. Incorporating single-cell data from tumors to parameterize these computational models provides deeper insights into prediction of clinical outcome in individual patients. Here, we integrate whole-exome sequencing and scRNA-seq data from Triple-Negative Breast Cancer patients to model neoantigen burden in tumor cells as input to a spatial Quantitative System Pharmacology model. The model comprises a four-compartmental Quantitative System Pharmacology sub-model to represent a whole patient and a spatial agent-based sub-model to represent tumor volumes at the cellular scale. We use the high-throughput single-cell data to model the role of antigen burden and heterogeneity relative to the tumor microenvironment composition on predicted immunotherapy response. We demonstrate how this integrated modeling and single-cell analysis framework can be used to relate neoantigen heterogeneity to immunotherapy treatment outcomes. Our results demonstrate feasibility of merging single-cell data to initialize cell states in multiscale computational models such as the spQSP for personalized prediction of clinical outcomes to immunotherapy.

Post-translational modifications of proteins play a crucial part in carcinogenesis. Phosphorylated peptides have shown to be presented by MHC class I molecules and recognised by cytotoxic T cells, making them a promising target for immunotherapy. Identification of phosphorylated MHC class I ligands has so far predominantly been done using bioinformatic tools trained on unmodified peptides. Only one tool, PhosMHCpred, has been developed specifically for the prediction of phosphorylated MHC class I ligands so far and this tool has been trained only on a limited number of alleles and provides a limited peptide length coverage (only including 9-mers).

Here we propose a method, termed NetMHCphosPan, for the prediction of MHC presented phosphopeptides. The method is trained using the NNAlign_MA framework, which allows incorporating mixed data types and information leverage between data sets resulting in a greatly improved MHC and peptide length coverage and an overall increased predictive power compared to PhosMHCpred. Motif deconvolution suggested a strong preference for phosphosites to be located in position 4 of the binding motif, and enrichment of proline at P5 and arginine at P1. The improved performance, driven by the extended length and allelic coverage, of NetMHCphosPan over current state-of-the-art methods, was further validated on a large benchmark data set independent from the model development.

In conclusion, we have confirmed the high power of NNAlign_MA for motif deconvolution of complex immuno-peptidomics data and have developed a novel method for prediction of MHC presented phosphopeptides with improved predictive power and a broader peptide length and MHC coverage compared to current state-of-the-art methods. The developed method is available at http://www.cbs.dtu.dk/services/NetMHCphosPan-1.0.