Identification of prognostic genes and gene sets for early-stage non-small cell lung cancer using bi-level selection methods

ECE2 is a prognostic biomarker associated with m6A modification and involved in immune infiltration of lung adenocarcinoma

BACKGROUND

The targeted therapy for lung cancer relies on prognostic genes and requires further research. No research has been conducted to determine the effect of endothelin-converting enzyme 2 (ECE2) in lung cancer.

METHODS

We analyzed the expression of ECE2 in lung adenocarcinoma (LUAD) and normal adjacent tissues and its relationship with clinicopathological characteristics from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus database (GEO). Immunohistochemical staining was used to further validate the findings. GO/KEGG enrichment analysis and gene set enrichment analysis (GSEA) of ECE2 co-expression were performed using R software. Data from TIMER, the GEPIA database, and TCGA were analyzed to determine the relationship between ECE2 expression and LUAD immune infiltration. To investigate the relationship between ECE2 expression levels and LUAD m6A modification, TCGA data and GEO data were analyzed.

RESULTS

ECE2 is highly expressed in various cancers including LUAD. ECE2 showed high accuracy in distinguishing tumor and normal sample results. The expression level of ECE2 in LUAD was significantly correlated with tumor stage and prognosis. GO/KEGG enrichment analysis showed that ECE2 was closely related to mitochondrial gene expression, ATPase activity and cell cycle. GSEA analysis showed that ECE2-related differential gene enrichment pathways were related to mitotic cell cycle, MYC pathway, PLK1 pathway, DNA methylation pathway, HIF1A pathway and Oxidative stress-induced cellular senescence. Analysis of the TIMER, GEPIA database, and TCGA datasets showed that ECE2 expression levels were significantly negatively correlated with B cells, CD4+ cells, M2 macrophages, neutrophils, and dendritic cells. TCGA and GEO datasets showed that ECE2 was significantly associated with m6A modification-related genes HNRNPC, IGF2BP1, IGF2BP3 and RBM1.

CONCLUSION

ECE2 is associated with m6A modification and immune infiltration and is a prognostic biomarker in LUAD.

The cox-filter method identifies respective subtype-specific lncRNA prognostic signatures for two human cancers

BACKGROUND

The most common histological subtypes of esophageal cancer are squamous cell carcinoma (ESCC) and adenocarcinoma (EAC). It has been demonstrated that non-marginal differences in gene expression and somatic alternation exist between these two subtypes; consequently, biomarkers that have prognostic values for them are expected to be distinct. In contrast, laryngeal squamous cell cancer (LSCC) has a better prognosis than hypopharyngeal squamous cell carcinoma (HSCC). Likewise, subtype-specific prognostic signatures may exist for LSCC and HSCC. Long non-coding RNAs (lncRNAs) hold promise for identifying prognostic signatures for a variety of cancers including esophageal cancer and head and neck squamous cell carcinoma (HNSCC).

METHODS

In this study, we applied a novel feature selection method capable of identifying specific prognostic signatures uniquely for each subtype - the Cox-filter method - to The Cancer Genome Atlas esophageal cancer and HSNCC RNA-Seq data, with the objectives of constructing subtype-specific prognostic lncRNA expression signatures for esophageal cancer and HNSCC.

RESULTS

By incorporating biological relevancy information, the lncRNA lists identified by the Cox-filter method were further refined. The resulting signatures include genes that are highly related to cancer, such as H19 and NEAT1, which possess perfect prognostic values for esophageal cancer and HNSCC, respectively.

CONCLUSIONS

The Cox-filter method is indeed a handy tool to identify subtype-specific prognostic lncRNA signatures. We anticipate the method will gain wider applications.

A DNA methylation signature to improve survival prediction of gastric cancer

Background

The current Union International Committee on Cancer or the American Joint Committee on Cancer TNM stage system has shown valuable but insufficient estimation for subsets of gastric cancer and prediction for prognosis patients. Thus, there is an urgent need to identify diagnostic, prognostic, and predictive biomarkers to improve patients’ outcomes. Our aim was to perform an integrative analysis on publicly available datasets to identify epigenetic changes that may play key role in the initiation and progression of gastric cancer, based on which we set to develop a DNA methylation signature to improve survival prediction of gastric cancer.

Results

A total of 340 methylation-related differentially expression genes (mrDEGs) were screened in gastric cancer patients from The Cancer Genome Atlas (TCGA) project. Pathway enrichment analysis revealed that they were involved in the biological process related to initiation and progression of gastric cancer. Based on the mrDEGs identified, we developed a DNA methylation signature consisting of ten gene members (SCNN1B, NFE2L3, CLDN2, RBPMS2, JPH2, GBP6, COL4A5, SMKR1, PPP1R14A, and ARL4D) according to their methylation β value. This innovative DNA methylation signature was associated with cancer recurrence, while it showed independence of cancer recurrence and TNM stage for survival prediction. Combination of this DNA methylation signature and TNM stage improved overall survival prediction in the receiver operating characteristic analysis. We also verified that two individual genes (PPP1R14A and SCNN1B) of the identified prognostic signature were regulated by promoter region methylation in a panel of gastric cell lines.

Conclusions

This study presents a powerful DNA methylation signature by performing analyses integrating multi-source data including transcriptome, methylome, and clinical outcome of gastric cancer patients from TCGA. The identified DNA methylation signature may be used to refine the current prognostic model and facilitate further stratification of patients in the future clinical trials. Further experimental studies are warranted to unveil the regulatory mechanism and functional role of all the individual genes of the DNA methylation signature. Also, clinical investigations in large GC patient cohorts are greatly needed to validate our findings.

Biomarker potential of repetitive-element transcriptome in lung cancer

Since repetitive elements (REs) account for nearly 53% of the human genome, profiling its transcription after an oncogenic change might help in the search for new biomarkers. Lung cancer was selected as target since it is the most frequent cause of cancer death. A bioinformatic workflow based on well-established bioinformatic tools (such as RepEnrich, RepBase, SAMTools, edgeR and DESeq2) has been developed to identify differentially expressed RNAs from REs. It was trained and tested with public RNA-seq data from matched sequencing of tumour and healthy lung tissues from the same patient to reveal differential expression within the RE transcriptome. Healthy lung tissues express a specific set of REs whose expression, after an oncogenic process, is strictly and specifically changed. Discrete sets of differentially expressed REs were found for lung adenocarcinoma, for small-cell lung cancer, and for both cancers. Differential expression affects more HERV-than LINE-derived REs and seems biased towards down-regulation in cancer cells. REs behaving consistently in all patients were tested in a different patient cohort to validate the proposed biomarkers. Down-regulation of AluYg6 and LTR18B was confirmed as potential lung cancer biomarkers, while up-regulation of HERVK11D-Int is specific for lung adenocarcinoma and up-regulation of UCON88 is specific for small cell lung cancer. Hence, the study of RE transcriptome might be considered another research target in cancer, making REs a promising source of lung cancer biomarkers.

Identification of Monotonically Differentially Expressed Genes for IFN-β-Treated Multiple Sclerosis Patients

Multiple sclerosis (MS) is a common neurological disability of the central nervous system. Immune-modulatory therapy with interferon-β (IFN-β) has been used as a first-line treatment to prevent relapses in MS patients. While the therapeutic mechanism of IFN-β has not been fully elucidated, the data of microarray experiments that collected longitudinal gene expression profiles to evaluate the long-term response of IFN-β treatment have been analyzed using statistical methods that were incapable of dealing with such data. In this study, the GeneRank method was applied to generate weighted gene expression values and the monotonically expressed genes (MEGs) for both IFN-β treatment responders and nonresponders were identified. The proposed procedure identified 13 MEGs for the responders and 2 MEGs for the nonresponders, most of which are biologically relevant to MS. Our work here provides some useful insight into the mechanism of IFN-β treatment for MS patients. A full understanding of the therapeutic mechanism will enable a more personalized treatment strategy possible.

Identification of monotonically differentially expressed genes for non-small cell lung cancer

Background

Monotonically expressed genes (MEGs) are genes whose expression values increase or decrease monotonically as a disease advances or time proceeds. Non-small cell lung cancer (NSCLC) is a multistage progression process resulting from genetic sequences mutations, the identification of MEGs for NSCLC is important.

Results

With the aid of a feature selection algorithm capable of identifying MEGs – the MFSelector method – two sets of potential MEGs were selected in this study: the MEGs across the different pathologic stages and the MEGs across the risk levels of death for the NSCLC patients at early stages. For the lung adenocarcinoma (AC) subtypes no statistically significant MEGs were identified across pathologic stages, however dozens of MEGs were identified across the risk levels of death. By contrast, for the squamous cell lung carcinoma (SCC) there were no statistically significant MEGs as either stage or risk level advanced.

Conclusions

The pathologic stage of non-small cell lung cancer patients at early stages has no prognostic value, making the identification of prognostic gene signatures for them more meaningful and highly desirable.

Incorporating Pathway Information into Feature Selection towards Better Performed Gene Signatures

To analyze gene expression data with sophisticated grouping structures and to extract hidden patterns from such data, feature selection is of critical importance. It is well known that genes do not function in isolation but rather work together within various metabolic, regulatory, and signaling pathways. If the biological knowledge contained within these pathways is taken into account, the resulting method is a pathway-based algorithm. Studies have demonstrated that a pathway-based method usually outperforms its gene-based counterpart in which no biological knowledge is considered. In this article, a pathway-based feature selection is firstly divided into three major categories, namely, pathway-level selection, bilevel selection, and pathway-guided gene selection. With bilevel selection methods being regarded as a special case of pathway-guided gene selection process, we discuss pathway-guided gene selection methods in detail and the importance of penalization in such methods. Last, we point out the potential utilizations of pathway-guided gene selection in one active research avenue, namely, to analyze longitudinal gene expression data. We believe this article provides valuable insights for computational biologists and biostatisticians so that they can make biology more computable.

Identification of subtype-specific prognostic signatures using Cox models with redundant gene elimination

Lung cancer (LC) is a leading cause of cancer-associated mortalities worldwide. Adenocarcinoma (AC) and squamous cell carcinoma (SCC) account for ~70% of all cases of LC. Since AC and SCC are two distinct diseases, their corresponding prognostic genes associated with patient survival time are expected to be different. To date, only a few studies have distinguished patients with good prognosis from those with poor prognosis for each specific subtype. In the present study, the Cox filter model, a feature selection algorithm that identifies subtype-specific prognostic genes to incorporate pathway information and eliminate redundant genes, was adopted. By applying the proposed model to data on non-small cell lung cancer (NSCLC), it was demonstrated that both redundant gene elimination and search space restriction can improve the predictive capacity and the model stability of resulting prognostic gene signatures. To conclude, a pre-filtering procedure that incorporates pathway information for screening likely irrelevant genes prior to complex downstream analysis is recommended. Furthermore, a feature selection algorithm that considers redundant gene elimination may be preferable to one without such a consideration.