Transcriptomic and proteomic intra-tumor heterogeneity of colorectal cancer varies depending on tumor location within the colorectum

Background Intra-tumor heterogeneity (ITH) of colorectal cancer (CRC) complicates molecular tumor classification, such as transcriptional subtyping. Differences in cellular states, biopsy cell composition, and tumor microenvironment may all lead to ITH. Here we analyze ITH at the transcriptomic and proteomic levels to ascertain whether subtype discordance between multiregional biopsies reflects relevant biological ITH or lack of classifier robustness. Further, we study the impact of tumor location on ITH. Methods Multiregional biopsies from stage II and III CRC tumors were analyzed by RNA sequencing (41 biopsies, 14 tumors) and multiplex immune protein analysis (89 biopsies, 29 tumors). CRC subtyping was performed using consensus molecular subtypes (CMS), CRC intrinsic subtypes (CRIS), and TUMOR types. ITH-scores and network maps were defined to determine the origin of heterogeneity. A validation cohort was used with one biopsy per tumor (162 tumors). Results Overall, inter-tumor transcriptional variation exceeded ITH, and subtyping calls were frequently concordant between multiregional biopsies. Still, some tumors had high transcriptional ITH and were classified discordantly. Subtyping of proximal MSS tumors were discordant for 50% of the tumors, this ITH was related to differences in the microenvironment. Subtyping of distal MSS tumors were less discordant, here the ITH was more cancer-cell related. The subtype discordancy reflected actual molecular ITH within the tumors. The relevance of the subtypes was reflected at protein level where several inflammation markers were significantly increased in immune related transcriptional subtypes, which was verified in an independent cohort (Wilcoxon rank sum test; p<0.05). Unsupervised hierarchical clustering of the protein data identified large ITH at protein level; as the multiregional biopsies clustered together for only 9 out of 29 tumors. Conclusion Our transcriptomic and proteomic analyses show that the tumor location along the colorectum influence the ITH of CRC, which again influence the concordance of subtyping.

Molecular tumor classification using DNA methylome analysis

Tumor classifiers based on molecular patterns promise to define and reliably classify tumor entities. The high tissue- and cell type-specificity of DNA methylation, as well as its high stability, makes DNA methylation an ideal choice for the development of tumor classifiers. Herein, we review existing tumor classifiers using DNA methylome analysis and will provide an overview on their emerging impact on cancer classification, the detection of novel cancer subentities and patient stratification with a focus on brain tumors, sarcomas and hematopoietic malignancies. Furthermore, we provide an outlook on the enormous potential of DNA methylome analysis to complement classical histopathological and genetic diagnostics, including the emerging field of epigenomic analysis in liquid biopsies.

WEIGHTED NEIGHBORHOOD CLASSIFIER FOR THE CLASSIFICATION OF IMBALANCED TUMOR DATASET

Machine learning is widely applied to gene expression profiles based molecular tumor classification, but sample imbalance problem is often overlooked. This paper proposed a subclass-weighted neighborhood classifier to address the imbalanced sample set problem and a novel neighborhood rough set model to select informative genes for classification performance improvement. Experiments on three publicly available tumor datasets demonstrated that the proposed method is obviously effective on imbalanced dataset with obscure boundary between two subtypes and informative gene selection and it can achieve higher cross-validation accuracy with much fewer tumor-related genes.