Increased Expression Together with the Gene Regulation Network of NEIL3 Predicts Poor Prognosis and Correlates with Immune Infiltrates in Hepatocellular Carcinoma

Background: Abnormal Nei endonuclease VIII-like 3 (NEIL3)expression is associated with carcinogenesis. Methods: We used sequencing data from the Cancer Genome Atlas database, analyzed NEIL3 expression, gene regulation networks and the correlation with immune infiltrates in hepatocellular carcinoma (HCC). Clinicopathologic characteristics associated with overall survival in TCGA patients using Cox regression and the Kaplan-Meier method. Gene Set Enrichment Analysis was performed using TCGA data set. LinkedOmics was used to identify differential gene expression with NEIL3 and to analyze Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathways. Gene enrichment analysis examined target networks of kinases and transcription factors.Correlations between NEIL3 expression and cancer immune infiltrates and immune gene markers were analyzed by TIMER and GEPIA. Results: We found that overexpressed NEIL3 predicted poor prognosis. Functional network analysis suggested that NEIL3 regulates the DNA replication and cell cycle signaling via pathways involving several cancer-related kinases and E2F Transcription Factor 1.NEIL3 was also found to be associated with the infiltration of several immune cells. Conclusions: Our results demonstrate that data mining efficiently reveals information about NEIL3 expression, potential regulatory networks and the relationship with immune infiltration in HCC, laying a foundation for further study of the role of NEIL3 in carcinogenesis.


Background
Hepatocellular carcinoma (HCC) is the second leading cause of cancer-related deaths worldwide, which affects about 800,000 cases every year [1]. Currently, HCC can be treated by surgical treatment and chemotherapy, but the mortality rate remains high [2]. The development of various targeted drugs has prolonged patient survival and made a revolutionary breakthrough in treating advanced HCC [3]. Nonetheless, no satisfactory efficacy is attained by the existing targeted drugs. In recent years, immune checkpoint inhibitors (ICIs), such as ipilimumab (the CTLA4 inhibitor) and nivolumab (the PD-1 inhibitor), demonstrate survival benefits for HCC, which reveals the close relationship between immune status and HCC progression [4]. The pathogenesis of HCC is extremely complex, which involves processes like cell cycle regulation and signal transduction, and reflects the function and interaction of multiple genes in multiple steps [5]. Typically, it may be possible to identify the novel drug targets of HCC through screening the gene networks for changes related to tumor genesis and progression.
Nei endonuclease VIII-like 3 (NEIL3), which serves as the DNA glycosylase, is up-regulated at the S phase and peaks at the late S/G2 phase; specifically, it preferentially repairs the oxidative lesions in telomere sequences in vitro [6]. NEIL3 efficiently excises the hydantoin lesions spiroiminodi-hydantoin (Sp) and guanidino-hydantoin (Gh) in single-stranded (ss) and double-stranded (ds) DNA; moreover, it also removes 5-hydroxy-2'-deoxycytidine (5OHC) and 5-hydroxy-2'-deoxyuridine (5OHU) from the ssDNA [7]. Unlike NEIL1 and NEIL2 that possess the β, δ-elimination activity, NEIL3 mainly incises the damaged DNA by means of β-elimination. In addition, the base excision and strand incision activities of NEIL3 were discordant, which indicates that NEIL3 mainly operates as a monofunctional DNA glycosylase. On the other hand, it is proposed that the NEIL3-dependent modulation of DNA methylation can regulate the proliferation of cardiac fibroblasts, thereby affecting lipid metabolism and extracellular matrix (ECM) modulation following myocardial infarction (MI) [8]. Klattenhoff et al. also demonstrated that, the loss of NEIL3 DNA glycosylase markedly increased the replicationassociated double strand breaks (DSBs) and improved the sensitivity to ATR inhibitor in glioblastoma cells [9]. Moreover, aberrant NEIL3 expression is observed in several cancers, and the median NEIL3 expression level is reported to be markedly positively correlated with the median somatic single and total mutation loads, as suggested by data on 13 cancer types from The Cancer Genome Atlas (TCGA) database [10]. The NEIL3 expression level in HCC is higher than that in normal hepatic tissues [11,12], however, the underlying molecular mechanism, together with its effect on HCC initiation and development, remains largely unclear so far. Moreover, the clinical relevance and prognostic significance of NEIL3 in HCC should be further explored.
In this study, NEIL3 expression and its correlation with the prognosis for HCC patients from databases (such as TCGA, HCCDB, UALCAN and Kaplan-Meier plotter) were comprehensively analyzed. Besides, the functional networks related to NEIL3 in HCC were also examined using the multi-dimensional analysis methods. Furthermore, the correlation of NEIL3 with the tumor-infiltrating immune cells (TIICs) was detected through the Tumor Immune Estimation Resource (TIMER). In a word, our results potentially revealed the novel targets and strategies for the diagnosis and treatment of HCC.

Materials And Methods 2.1. Acquisition of clinical samples and patient data for RNA-sequencing
The transcriptomic RNA-sequencing data from HCC samples were downloaded from the TCGA data portal (https://cancergenome.nih.gov/), which covered data from 374 primary HCC tissues and 50 non-carcinoma tissues. The NEIL3 expression levels were then extracted. At the same time, the clinicopathological characteristics for these patients, including age, gender, pathological stage (grade), tumor stage, T stage, lymph node metastasis (LNM) and distant metastasis (DM), were also downloaded and collected. In addition, HCCDB (http://lifeome.net/database/hccdb/about.html), the one-stop online resource that contains 15 public HCC expression datasets involving up to about 4000 clinical samples for exploring the HCC gene expression, was also adopted for more extensive validation [13]. Supplementary Table 1 presents the correspondence between the HCCDB database and the Gene Expression Omnibus (GEO) database.

Survival curve plotting and clinical correlation analysis
Clinical data downloaded from TCGA data portal were collected to extract the survival time, and the primary endpoint was deemed as death. Further, the survival curve was plotted using the survival package of R software. Then, the relationship of clinical parameters with NEIL3 expression was also examined to reveal the underlying clinical significance of NEIL3. Noteworthily, the Kaplan-Meier plotter, which can assess the impacts of 54 k genes on the survival for 21 cancer types (http://kmplot.com/analysis/), was also utilized to assist in estimating the effect of NEIL3 on survival [14,15]. The largest datasets in the Kaplan-Meier plotter include breast (n = 6,234), ovarian (n = 2,190), lung (n = 3,452), and gastric (n = 1,440) cancers; whereas the miRNA subsystems cover 11 k samples from 20 cancer types. This system has contained the gene chip and RNA-seq data sources for databases like GEO, EGA, and TCGA. In this study, the hazard ratio (HR) with the corresponding 95% confidence intervals (CI), and the log-rank P-value were also calculated. In view of the small number of patients at stage 4, patients at stage 3 and 4 were also adopted for analysis.
Moreover, UALCAN, the interactive web-portal that uses TCGA level 3 RNA-seq and clinical data from 31 cancer types for in-depth analyses on TCGA gene expression data, was also employed to assist in analysis [16]. Typically, the portal possesses the user-friendly feature of allowing for analysis on the relative expression of a query gene(s) across tumor and normal samples, as well as among various tumor sub-groups according to individual cancer stages, tumor grade or other clinicopathological features. UALCAN is publicly available at http://ualcan.path.uab.edu.

Gene set enrichment analysis (GSEA)
GSEA is a computational approach, which determines whether an a priori defined set of genes shows statistically significant and concordant differences between two biological states [17]. In our study, an ordered list of all genes was generated by GSEA at first based on their correlations with NEIL3 expression; afterwards, GSEA was conducted to illustrate the significant survival difference observed between the high and low NEIL3 expression groups. Typically, the gene set permutations were performed for 1000 times in each analysis, with NEIL3 expression being used as a phenotype label. In addition, the nominal p-value and the normalized enrichment score (NES) were adopted to sort the pathways enriched in each phenotype. Moreover, multiple GSEA was also carried out from a comprehensive perspective.

LinkedOmics analysis
The LinkedOmics database (http://www.linkedomics.org/login.php) is a Web-based platform for analyzing 32 TCGA cancer-associated multi-dimensional datasets [18]. In this study, the LinkFinder module of LinkedOmics was utilized to examine the differentially expressed genes (DEGS) correlated with NEIL3 in TCGA LIHC cohort (n = 371). The Spearman correlation coefficient was used to statistically analyzed the results, and the statistical plots for individual genes were also created. All results were graphically presented in the forms of volcano plots, heat maps or scatter plots. Notably, the LinkInterpreter module of LinkedOmics carries out pathway and network analyses of the DEGs. In this study, data obtained from the LinkFinder results were signed and ranked, and GSEA was employed for Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, kinase-target enrichment, and transcription factor-target enrichment analyses. Notably, the final network analyses were conducted based on the Molecular Signatures Database (MSigDB). FDR of < 0.05 was selected as the rank criterion, and 500 simulations were performed.

TIMER database analysis
TIMER is a web resource to systematically evaluate the clinical impacts of different immunocytes on various cancer types, which covers 10,879 samples across 32 cancer types from TCGA to estimate the abundances of six tumor-infiltrating immunocyte subtypes, including B cells, CD4 T cells, CD8 T cells, macrophages, neutrophils, and dendritic cells (https://cistrome.shinyapps.io/timer/) [19]. This database applies the previously published statistical method deconvolution to predict the abundances of tumor-infiltrating immune cells (TIICs) based on the gene expression profiles [20]. This study analyzed the correlation of NEIL3 expression with the abundances of immune infiltrates in HCC, including B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages, and dendritic cells, using the gene modules. Besides, the gene expression levels against tumor purity were also displayed [21]. Tregs, and exhausted T cells. These gene markers had been reported in previous studies [22][23][24][25][26].
NEIL3 expression was used as the x-axis with gene symbols, whereas related marker genes were displayed on the y-axis as gene symbols. The gene expression level was expressed in the form of log2 RSEM.

Gene correlation analysis in GEPIA
The online database Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancerpku.cn/index.html) is an interactive web, which covers 9,736 tumors and 8,587 normal samples from TCGA and the Genotype-Tissue Expression (GTEx) projects, and analyses the RNA sequencing data [27]. In this study, GEPIA was adopted to further confirm genes with significant correlation in TIMER. Additionally, gene expression correlation analysis was carried out on the given sets of TCGA expression profile data. The correlation coefficient was determined according to the Spearman method. NEIL3 was used as the x-axis, while the other genes of interest were presented on the y-axis, and both tumor and normal tissue datasets were used for analysis.

Statistical analysis
The R (v.3.6.1) package was adopted for all statistical analyses. The relationships of clinicopathological features with NEIL3 expression were examined through the Wilcoxon signed-rank test and logistic regression. Thereafter, the clinicopathological characteristics associated with the OS for TCGA patients were determined using Cox regression and the Kaplan-Meier method. Further, the effects of NEIL3 expression on patient survival along with other clinical characteristics (such as age, gender, stage, grade, T stage, LNM status, and DM status) were compared using the multivariate Cox analysis. The threshold NEIL3 expression was determined by its median level. Besides, the results of Kaplan-Meier plots, UALCAN and GEPIA were presented as HRs and P-values or Cox P-values upon the log-rank test. Moreover, Spearman's correlation and statistical significance was conducted to evaluate the correlations of gene expression. P-values of < 0.05 were considered as statistically significant.

NEIL3 mRNA expression in HCC and other human cancers
To determine the differences in NEIL3 expression between HCC and normal tissues, the differential scatter plot and paired plot were obtained ( Fig. 1A and 1B). Also, NEIL3 expression among different GEO datasets was analyzed using the HCCDB database. Furthermore, NEIL3 mRNA expression within various cancer types and matched normal tissues were measured based on the TIMER database, so as to reveal the aberrant NEIL3 expression from a holistic perspective. Our analysis suggested that NEIL3 expression was up-regulated in HCC among different datasets (Fig. 1C), as well as in bladder, breast, biliary, colorectal, esophageal, gastric, kidney, lung, prostatic and rectal cancers compared with that in normal tissues (Fig. 1D).

Survival outcomes and multivariate analysis
The Kaplan-Meier survival analysis in Fig. 2A suggested that HCC with high NEIL3 expression was associated with worse prognosis than that with low NEIL3 expression in terms of TCGA (p < 0.001).
Similar results were also obtained from the analysis of HCC samples in the International Cancer Genome Consortium (ICGC) HCC cohort (Fig. 2B). The values of area under the receiver operating characteristic (ROC) curve (AUC) for 1-, 3-and 5-year OS of TCGA patients were 0.596, 0.663 and 0.737, respectively (Fig. 2C). In addition, to better illustrate the relevance and underlying mechanisms of NEIL3 expression in HCC, the relationships of NEIL3 expression with clinical characteristics and survival prognosis for HCC patients were analyzed in the Kaplan-Meier plotter database (Table 1). According to our results, NEIL3 over-expression was related to worse OS in male and female patients, white and Asian patients, together with those with different clinical stages and T stages upon the American Joint Committee on Cancer (AJCC), none or micro-vascular invasion status, sorafenib treatment as well as both risk factors (P < 0.05). Specifically, the effects of high NEIL3 expression on OS were particularly significant in Asian (HR = 6.54, P < 0.001) and G1 (HR = 4.18, P < 0.01) patients. Besides, clinical stage and T stage were also associated with the poor patient survival (Fig. 4A).

Associations between NEIL3 expression and clinicopathological variables
Altogether 371 HCC samples with available NEIL3 expression data stratified based on multiple patient characteristics were analyzed from TCGA. Figure 3A consistently suggested the high transcription of NEIL3. As shown in Fig. 3B-3G, the transcription levels of NEIL3 were notably higher in HCC patients than in healthy people upon subgroup analyses stratified according to gender, age, weight and obese, ethnicity, disease stage, tumor grade and nodal metastasis status. Additionally, the univariate analysis using logistic regression revealed that NEIL3 expression, which served as a categorical dependent variable (based on median expression level of 2.5), was linked with the clinicopathological characteristics predicting dismal prognosis ( -5C); in addition, the complement and coagulation cascades, primary bile acid biosynthesis, valine leucine and isoleucine degradation were primarily enriched in low ANLN expression group (FDR < 0.05, NOM < 0.005) (Fig. 5D-5F). The markedly enriched signaling pathways related to cancer and immunity, including the P53 pathway, the T cell receptor pathway, the mTOR pathway, the ERBB pathway, the Notch signaling pathways and Pathways in cancer were also differentially enriched in the high NEIL3 expression phenotype (Supplementary Table 2). Afterwards, multiple GSEA analysis was also conducted to further demonstrate the whole picture of the signaling pathways related to NEIL3 expression in HCC (Fig. 5H).

GO and KEGG pathway analyses of the co-expressed genes correlated with NEIL3 in HCC
The Function module of LinkedOmics was adopted to analyze the mRNA sequencing data from 371 LIHC patients in TCGA. As shown in the volcano plot ( aurora kinase A (AURKA) ( Table 3). In addition, the transcription factor-target network was mainly related to the E2F Transcription Factor (E2F) family, including E2F1_Q6, E2F1_Q4 and E2F4DP1_01.
The protein-protein interaction (PPI) network constructed by GeneMANIA revealed correlations among genes for kinases PLK and TF E2F1_Q6; besides, the gene sets enriched for them were mainly responsible for nuclear division, organelle fission, mitosis and DNA replication. (Fig. 8 and Supplementary Fig. 2). Table 3 The Kinase, miRNA and transcription factor-target networks of NEIL3 in hepatocellular carcinoma (LinkedOmics) Enriched category Gene set leadingEdgeNum FDR Kinase Target  Kinase_PLK1  33  0  Kinase_CDK1  84  0  Kinase_AURKB  26  0  Kinase_CHEK1  41  0  Kinase_AURKA  16  0  Transcription factor  macrophages (r = 0.296, p = 2.46e -08) and neutrophils (r = 0.273, p = 2.71e -07) (Fig. 9). Moreover, functional annotation was further carried out by means of GSEA, and the results suggested that the differentially expressed genes between these two groups were enriched in the immunological signature gene sets (c7. All. V7.0. symbol) (Supplementary Table 3). Taken together, these findings suggested that NEIL3 might play a specific role in the immune infiltration in HCC, especially for B cells, T cells and dendritic cells.

Correlation analysis between NEIL3 expression and immune marker sets
To further probe into the relationships between NEIL3 and the diverse immune infiltrating cells, the correlations of NEIL3 with the immune marker sets of various HCC immunocytes were explored in the TIMER and GEPIA databases. As suggested by our results, NEIL3 expression was correlated with the immune marker genes of various immune cells, including CD8 + T cells, T cells (general), B cells, monocytes, TAMs, M1 and M2 macrophages, and neutrophils in HCC (Table 4). Also, the diverse functional T cells were also analyzed, such as Th1 cells, Th2 cells, Tfh cells, Th17 cells, Tregs, and the exhausted T cells. Results of correlation analysis adjusted by purity revealed that, the NEIL3 expression level was significantly correlated with the most immune marker sets of various immune cells and the diverse functional T cells in HCC. Moreover, those markers associated with T cells were selected to analyze their correlations with NEIL3 expression in the GEPIA database. According to our results, NEIL3 expression was markedly correlated with the markers related to T cells in HCC, but not with most of them in normal tissues ( Table 5), indicating that NEIL3 might be involved in regulating T cell function.

Discussion
The genetic abnormalities of NEIL3, a DNA glycosylase that initiates base excision repair by hydrolyzing the N-glycosidic bond and releasing the damaged base, are reported to be correlated with hepatocarcinogenesis [28]. Recent study also demonstrates that the Neil3 gene expression levels are highly related to HCC recurrence in patients receiving hepatectomy, which suggests that NEIL3 may serve as a prognostic marker for HCC [11]. To the best of our knowledge, NEIL3 expression, together with its potential prognostic impact on HCC, has not yet been explored, and the potential role of NEIL3 in HCC is the focus in this study. To further investigate the functions of NEIL3 in HCC, GSEA was implemented using TCGA data, and then the differentially enriched pathways in the high NEIL3 expression phenotype as well as those linked to cancer and immunity, were selected. Among them, Base excision repair scored 2.17 on NES, and the polymorphisms of base-excision repair genes are reported to shape the development of hepatitis C virus (HCV) infection, liver cirrhosis and HCC, which predicts the OS [29,30]. Besides, plenty of studies have elaborated that, the Notch signaling that may induce epithelial-mesenchymal transition (EMT) and mutation or loss of P53, the ErbB pathway that regulates the fundamental cellular processes, and the mTOR signaling that determines cell fate at different levels, play crucial roles in hepatocarcinogenesis [31][32][33]. Nevertheless, this study was the first to propose that, NEIL3 might be involved in these pathways in the HCC setting, suggesting that NEIL3 might serve as a potential prognostic marker and therapeutic target of HCC.
To further probe the corrections between NEIL3 and other genes in HCC, the interactions between NEIL3 and other functional partners were investigated in the LinkedOmics database. Our results revealed that, there were plenty of genes co-expressed with NEIL3 in LIHC, among which, CENPE, a spindle checkpoint protein, exhibited the highest Spearman's correlation of 0.774. Of interest, CENPE has been identified as a tumor suppressor gene, and its decreased expression may contribute to HCC development [34]. However, recent study documents that CENPE is remarkably up-regulated in HCC [35], consistent with our results on CENPE in TCGA. Therefore, CENPE was speculated to play a dual role in HCC progression. NCAPG, which ranks the second in our list, has shown its effect on HCC cell proliferation and migration, and it is evidently correlated with HCC recurrence, metastasis, differentiation and TNM stage [36]. Additionally, Hu et al. demonstrated that the over-expression of KIF4A, which was also highly correlated with NEIL3, was regulated by forkhead box M1 (FOXM1); besides, it had a bearing on HCC recurrence and progression, and might serve as a promising prognostic marker [37]. Furthermore, enrichment analyses on target gene sets using GSEA helps to reveal the important networks of target kinases and transcription factors (TFs). Our results suggested that the functional network of NEIL3 mainly participated in chromosome segregation, spindle organization, cell cycle regulation, and DNA replication. Interestingly, NEIL3 is also proposed in previous study to be involved in the cell signaling pathway that senses oxidative stress (OS), and contributes to the recruitment of TFs as well as the introduction of chromatin modifications, thereby resulting in the fine tuning of cellular processes for promoting cell survival and successful cell differentiation [38]. Moreover, among the highly NEIL3-related kinases investigated in this study, PLK1, one of the conserved serine/threonine kinase family members involved in multiple mitotic processes (including functional maturation of centrosomes, establishment of the bipolar spindle, chromosome segregation and response to DNA damage), was suggested in previous study to be a factor that determined the SN38 sensitivity of tumor cells, and its over-expression was inversely correlated with the survival rate of HCC [39]. Besides, other kinases, such as CDK1 and AURKB, which are involved in our network, also regulate genomic stability, mitosis and the cell cycle [40,41]. Among the TFs targeted by NEIL3, E2F1 represents a key link in the cell cycle regulation network, and its aberrant expression participates in HCC occurrence and development, which also predicts the unfavorable patient prognosis [42]. As reported, E2F1, which is tightly linked to cell division cycle associated 5 (CDCA5), Sirtuin 5 (SIRT5) and other important molecules, may serve as a vital anti-apoptotic factor in liver cancer due to its ability to offset c-Myc-driven apoptosis [43,44]. As suggested in previous literature, the repressive complex, which includes the component of E2F4, is relieved by CDK upon reentry into the cell cycle; thereafter, E2F1-Sp1, which is in the form of complex E2F1-NF-Y, can be recruited to the promoter [45]. Therefore, it was deduced in this study that NEIL3 might regulate this process. Therefore, our analyses suggested that PLK, CDK, E2F1, and E2E4 were all the concernful targets of NEIL3, and that NEIL3 acted through these factors to regulate the cell cycle and proliferation capacity of HCC. Nonetheless, further studies are warranted to test this hypothesis.
Another important discovery of this study was that, NEIL3 expression was correlated with the diverse immune infiltration levels in HCC. As demonstrated by our results, the infiltration levels of which is transiently associated with the phosphatase SHP2 [46]. Typically, NEIL3 may exert its own strength to participate in this system. CCR8, together with CCR10, is the chemokine receptor responsible for Treg cell migration to the tumor microenvironment (TME) [47], and high NEIL3 expression is also proved to be linked with this molecule. On the other hand, TIM-3, a critical surface protein on the exhausted T cells [48], Nonetheless, some limitations should be noted in this study, which should be taken into consideration when interpreting our results. Firstly, transcriptomic analysis only reflected some aspects of genes, rather than the global alterations. Secondly, there were few stage 4 patients in the LIHC samples, since most HCC patients were first diagnosed at the advanced stage with dismal prognosis. Finally, our results were not validated via another independent cohort, which was also a limitation in this study, and the reliability of our molecular results was still challenged due to the lack of experiments in vitro or in vivo.
In conclusion, NEIL3, which may be positively co-expressed with CEPNE, NCAPG and KIF14, can serve as a potential prognostic molecular marker for the poor survival of HCC. Moreover, the base excision repair, the P53 pathway, the T cell receptor pathway, the mTOR pathway, the ERBB pathway and the Notch signaling may be the key pathways that are regulated by NEIL3; whereas PLK, CDK, E2F1_Q6 and E2F1_Q4 in EC may be targeted by NEIL3, and are involved in cell cycle regulation and DNA replication. Additionally, NEIL3 is also found to be associated with the infiltration of immune cells, especially for the functions of certain T cells related to HCC. Further experimental validation should be performed to confirm the biological impact of NEIL3.

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this article and its supplementary information files.
the network edge indicate the bioinformatics methods applied: co-expression, website prediction, pathway, physical interactions, shared protein domains, genetic interaction and co-localization. The different colors for the network nodes indicate the biological functions of the set of enrichment genes. Figure 1 Comparison of NEIL3 expression between normal tissue and hepatocellular carcinoma. (A) and (B) indicates that NEIL3 expression levels in hepatocellular carcinoma is significantly higher than that in normal tissues. (C) Human NEIL3 expression levels in different tumor types from TCGA database were determined by TIMER (*P < 0.05, **P < 0.01, ***P < 0.001).

Figures
(D) Expression of NEIL3 in different hepatocellular carcinoma datasets.                Correlation of NEIL3 expression with immune infiltration level in hepatocellular carcinoma.
Expression has significant positive correlations with infiltrating levels of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells, other than tumor purity.

Figure 9
Correlation of NEIL3 expression with immune infiltration level in hepatocellular carcinoma.
Expression has significant positive correlations with infiltrating levels of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells, other than tumor purity.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.