Skip to main content
PMC home page
IET Systems Biology logoLink to IET Systems Biology
. 2022 Nov 10;16(6):220–229. doi: 10.1049/syb2.12054

Genes associated with diagnosis and prognosis of Burkitt lymphoma

Albert Doughan 1, Samson Pandam Salifu 1,2,✉
PMCID: PMC9675412  PMID: 36354023

Abstract

Burkitt lymphoma (BL) is one of the most aggressive forms of non‐Hodgkin's lymphomas that affect children and young adults. The expression of genes and other molecular markers during carcinogenesis can be the basis for diagnosis, prognosis and the design of new and effective drugs for the management of cancers. The aim of this study was to identify genes that can serve as prognostic and therapeutic targets for BL. We analysed RNA‐seq data of BL transcriptome sequencing projects in Africa using standard RNA‐seq analyses pipeline. We performed pathway enrichment analyses, protein–protein interaction networks, gene co‐expression and survival analyses. Gene and pathway enrichment analyses showed that the differentially expressed genes are involved in tube development, signalling receptor binding, viral protein interaction, cell migration, external stimuli response, serine hydrolase activity and PI3K‐Akt signalling pathway. Protein–protein interaction network analyses revealed the genes to be highly interconnected, whereas module analyses revealed 25 genes to possess the highest interaction score. Overall survival analyses delineated six genes (ADAMTSL4, SEMA5B, ADAMTS15, THBS2, SPON1 and THBS1) that can serve as biomarkers for prognosis for BL management.

Keywords: Africa, biomarker, Burkitt lymphoma, differential expression analyses, hub genes

graphic file with name SYB2-16-220-g003.jpg

Analysing gene expression data of Burkitt lymphoma transcriptome sequencing from African patients revealed six biomarkers (ADAMTSL4, SEMA5B, ADAMTS15, THBS2, SPON1 and THBS1) as possible indicators for diagnostic and prognostic targets towards BL management.

1. INTRODUCTION

Burkitt lymphoma (BL) is one of the most aggressive forms of non‐Hodgkin's lymphomas that affect children and young adults. Overall, BL accounts for 1%–5% of all non‐Hodgkin's lymphomas and clinically presents as a conspicuous accumulation of tissues in the cheek and jaws of its victims [1]. The pattern and distribution of BL vary depending on age, sex and geographical location, suggesting the roles of both genetic and environmental factors in the development of the disease. BL is more common in males than in females with a ratio of 4:1 [2], accounts for 36% of all childhood cancers (2–16 years), 70% of childhood lymphomas and 5% of lymphomas for both adults and children [3]. In equatorial Africa, BL is usually associated with Plasmodium falciparum and Epstein‐Barr virus infections, where the latter is seen in 90% of BL cases [4]. Although BL can be treated if detected early, identifying reliable biomarkers has the added advantage of outcome assessment.

Identifying biomarkers associated with BL is important as it gives detailed information about the likely outcome of a treatment regimen. Panea et al. [5] identified BCL7A and BCL6 as potential predictive biomarkers in a Kenyan BL cohort. Kaymaz et al. [6] also found four members of the Proteasome 20S Subunit Beta (PSMB9, PSMB10, PSMB8 and PSMB2) to be associated with BL progression. The signature translocation event in BL involving the MYC and its deregulation has also been revealed by Kaymaz et al. [6], Abate et al. [7] and Panea et al. [5] through RNA‐seq analyses. LEF1 has been reported as a biomarker for assessing prognosis in multiple human cancers, including BL [8]. However, the following gaps persist in the current BL biomarkers; [1] there is no established consensus regarding the identified biomarkers implicated in BL progression across the various geographic locations; [2] there is limited information about the interactions among the identified biomarkers; and [3] there is inadequate information on how the biomarkers compare with biomarkers from other human cancers. The present study leverages these gaps by analysing a pool of RNA‐seq data from different geographical locations in Africa. This work provides a broader picture of the genetic landscape of BL in Africa, and the identified biomarkers could serve as a more representative biomarker associated with BL progression. Ultimately, the biomarkers could be further explored as potential targets for the diagnostic and prognostic makers for management of BL.

2. MATERIALS AND METHODS

2.1. Data acquisition

In this study, we downloaded and analysed RNA‐seq data obtained from National Centre for Biotechnology Information Short Read Archive (NCBI‐SRA) that met our set criteria of [1] the data being generated from African patients with BL within the last 10 years, [2] diagnosis of BL being conducted and confirmed by two experienced oncology pathologists and [3] data is available as paired‐end sequences. We excluded patients with either immunodeficient BL or sporadic BL. With these criteria, we arrived at three different sets of RNA‐seq data generated by Abate et al. [7], Lombardo et al. [9] and Kaymaz et al. [6]. The raw RNA‐seq data was downloaded from the NCBI‐SRA under the Accession numbers SRP062178, SRP099346 and SRP009316. Overall, 100 RNA‐seq data were included in the analysis; 50 cases and 50 control groups. Figure 1 provides an overview of the steps undertaken in this study.

FIGURE 1.

FIGURE 1

Open in a new tab

A diagrammatic representation of the various analyses performed in this study.

2.2. Data pre‐processing, trimming and alignment

The raw RNA‐seq datasets in the FASTQ format were evaluated for gastric cancer (GC)‐content, k‐mer distribution, sequence counts, quality scores, base quality, adaptor content and sequence duplication levels with FastQC v0.11.8 [10]. MultiQC v1.9 [11] was used to summarise all the reports generated by FastQC. Trimmomatic v0.39 [12] was used to trim off adaptors, low‐quality bases and short sequences. Sequences that passed the quality trimming stage were mapped to the human reference genome (GRCh38) using STAR v2.7.7a [13]. featureCounts v1.6.1 was used for gene quantification.

2.3. Differential expression analysis (DEA) and gene ontology (GO) analyses

Differential expression analysis (DEA) was performed within the R statistical tool's environment using DESeq2 under the default settings. Genes with adjusted p‐values less than 0.05 were considered as differentially expressed between the two conditions (cases vs. controls). Gene ontology (GO) analysis was performed using the ShinyGO online platform [14]. GO achieves high throughput annotation of genes based on the biological process, molecular function, cellular component and biological pathway analyses. Pathways and biological processes (BP) with adjusted p values less than 0.05 were considered statistically significant.

2.4. Protein–protein interaction (PPI) network

STRING [15] was used to retrieve the complex interactions between the genes and visualised with Cytoscape [16]. MCODE [17], a Cytoscape‐based software that topologically clusters a network to locate areas of dense connectedness was used to identify hub genes.

2.5. Hub genes expression in tumours and survival analysis

The expression of the hub genes in other human cancers was assessed using the Gene Expression Profiling Interactive Analysis 2 (GEPIA2) online tool [18]. GEPIA2 is an interactive resource that enables researchers to gather valuable information from genes using gene expression data from The Cancer Genome Atlas (TCGA) and Genotype‐Tissue Expression (GTEx) projects. Patient survival analysis was performed using the GEPIA2.

3. RESULTS

3.1. Selecting differentially expressed genes

In the present study, we analysed 100 RNA‐seq data generated from patients with BL and geographically matched healthy controls. Compared with the normal samples, 21,286 genes were differentially expressed (Figure 2). Setting the fold change threshold to ±2, a total of 6314 genes were arrived at, of which 4291 and 2023 were down‐ and up‐regulated respectively.

FIGURE 2.

FIGURE 2

Open in a new tab

A volcano plot of the RNA seq results. In all 6314 differentially expressed genes were determined. The blue and yellow dots represent down‐regulated and up‐regulated genes respectively. The black dots are not differentially expressed genes, and all dots below the red line are non‐significant genes.

3.2. Gene ontology (GO) and pathway enrichment analyses

We used the ShinyGO online tool to establish the GO classifications and pathways for the detected DEGs to elucidate their functions. The GO terms were grouped as BP, molecular functions (MF) and cellular components (CC), together with the Kyoto Encyclopaedia of Genes and Genomes (KEGG) was used to map the molecular pathways (Figure 3). The results showed that the DEGs were primarily enriched in BP, such as tube development (4.1 × 10−56), circulatory system development (1.7 × 10−55), anatomical structure formation involved in morphogenesis (2 × 10−53), cell migration (1.8 × 10−72) and biological adhesion (2 × 10−69). The DEGs for MF were enriched in signalling receptor binding (3.8 × 10−32), extracellular matrix binding (6.5 × 10−14), extracellular matrix structural constituent (3.8 × 10−32), cytokine receptor activity (2.8 × 10−13), integrin binding (1.9 × 10−16) etc. The results from the CC analysis showed that the genes are involved in collagen trimer (3.8 × 10−17), collagen‐containing extracellular matrix (1.9 × 10−73), extracellular matrix (1.6 × 10−82), external encapsulating structure (1.6 × 10−82), receptor complex (6.5 × 10−28) etc. The DEGs were significantly enriched in the following pathways; Malaria (3 × 10−11), ECM‐receptor interaction (2.0 × 10−14), viral protein interaction with cytokine and cytokine receptor (4.9 × 10−16), haematopoietic cell lineage (3.7 × 10−9), protein digestion and absorption (8.8 × 10−11) etc. (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

Gene ontology (GO) and pathway enrichment analysis of the DEGs. (a) GO biological process, (b) GO cellular component, (c) GO molecular function and (d) Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway. The dot size represents the number of DEGs implicated in the respective GO term. The colour intensity represents the significance level; red is more significant than blue.

3.3. PPI network analyses and module selection

Cytoscape was used to generate the protein‐protein interaction (PPI) network plot of all the DEGs (Figure 4a). The most highly interconnected regions were selected using the MCODE plugin in Cytoscape. Based on the interaction score, the resulting PPI networks were ranked. Figure 3b shows the network with the highest interaction score (24.92) among all the 78 networks. It contained 25 nodes (genes) and 299 edges (interactions). Through GO and KEGG analysis, the 25 hub genes were found to be enriched in metalloendopeptidase activity, extracellular matrix organisation, extracellular matrix, Malaria etc. (Table 1).

FIGURE 4.

FIGURE 4

Open in a new tab

Protein–protein interaction (PPI) network of the differentially expressed genes. (a) Interactions between all the 6314 genes. (b) Interactions between the highly interconnected genes. Circles and dark lines represent genes and interactions respectively.

TABLE 1.

Gene ontology and KEGG enrichment analyses of the 25 hub genes

GO category Term ID Description Adjusted p‐value No. of genes
MF GO:0004222 Metalloendopeptidase activity 5.74 × 10−31 17
MF GO:0008237 Metallopeptidase activity 2.52 × 10−26 17
MF GO:0004175 Endopeptidase activity 5.34 × 10−20 17
MF GO:0008233 Peptidase activity 3.26 × 10−17 17
MF GO:0016787 Hydrolase activity 6.32 × 10−08 17
BP GO:0030198 Extracellular matrix organisation 6.17 × 10−20 17
BP GO:0019538 Protein metabolic process 8.15 × 10−4 20
BP GO:0045229 External encapsulating structure 6.98 × 10−20 17
BP GO:0006508 Proteolysis 6.53 × 10−12 19
BP GO:0016043 Cellular component organisation 7.5 × 10−4 21
CC GO:0031012 Extracellular matrix 2.01 × 10−24 20
CC GO:0030312 External encapsulating structure 2.09 × 10−24 20
CC GO:0005576 Extracellular region 2.25 × 10−10 23
CC GO:0071944 Cell periphery 3.14 × 10−6 22
CC GO:0005788 Endoplasmic reticulum lumen 1.10 × 10−2 5
KEGG KEGG:05144 Malaria 1.69 × 10−2 2
Open in a new tab

Abbreviations: BP, Biological process; CC, Cellular component; GO, Gene ontology; KEGG, Kyoto Encyclopaedia of Genes and Genomes; MF, Molecular function.

3.4. Hub gene verification and co‐expression analysis

We verified the hub genes using GEPIA2 with data from the TCGA and GTEx projects. Figure 5 depicts the levels of expression of the individual genes within the various human cancer types. Among the cancer types, Acute Myeloid Leukemia and Diffuse Large B‐Cell (DLBC) lymphoma were the closest to BL, and in both cases, there was a relatively low expression of the hub genes.

FIGURE 5.

FIGURE 5

Open in a new tab

Hub gene expression in multiple human cancers. Deeper colours represent the overexpression of the respective genes. The cancer type abbreviations can be found here: https://gdc.cancer.gov/resources‐tcga‐users/tcga‐code‐tables/tcga‐study‐abbreviations.

Gene co‐expression analysis with STRING also revealed the interactions between the hub genes (Figure 6). RNA expression patterns and protein co‐expression values from the ProteomeHD database were utilised to construct the confidence scores used to generate the relationships. Figure 6 shows that more than half of the genes positively interact.

FIGURE 6.

FIGURE 6

Open in a new tab

Co‐expression analysis of the 25 hub genes. The deeper the colour, the stronger the association.

3.5. Hub gene survival analyses

The Kaplan‐Meier plotter in GEPIA2 was used to assess the predictive significance of 25 hub genes in the PPI network. The high and low expression of each gene was used to evaluate the overall survival (OS) of BL patients. From Figure 7, the overexpression of ADAMTSL4 and THBS1 was negatively associated with OS in BL patients. Low expression of SEMA5B, ADAMTS15, THBS2 and SPON1 was associated with a better OS in BL patients. The rest of the 25 hub genes had no significant correlation (adjusted p > 0.05) in patients with BL.

FIGURE 7.

FIGURE 7

Open in a new tab

Kaplan–Meier overall survival (OS) analysis of the most significant hub genes. The x and y axes represent the time (in months) and percent survival, respectively. (a) ADAMTSL4, (b) SEMA5B, (c) ADAMTS15, (d) THBS2, (e) SPON1 and (f) THBS1.

4. DISCUSSION

BL is one of the leading causes of facial disfigurations and deaths among children in sub‐Saharan Africa and Northeastern Asia. BL has substantial socio‐economic implications for the patients, their families, and the society. Although the genes implicated in the progression of BL have been extensively explored, research on specific prognostic, diagnostic and therapeutic markers for BL is limited. This study is novel, as it explores common biomarkers that could be used in the management of BL across Africa. In the present study, we pooled and analysed RNA‐seq data from several African BL sequencing projects. Our objective was to delineate a common gene expression pattern for BL patients across Africa and explore genes that can be diagnostic and prognostic markers for BL management.

We ran the data through a standard RNA‐seq analyses pipeline, using DESeq2 as the DEA tool. We found 6314 genes to be significantly differentially expressed (adjusted p‐value less than 0.05 and fold change of ±2). The downregulated genes were 4291, whereas 2023 were upregulated. Gene and pathway enrichment analyses showed that the DEGs were primarily implicated in tube development, signalling receptor binding, viral protein interaction, protein‐protein digestion and cell migration, response to external stimulus, serine hydrolase activity and PI3K‐Akt signalling pathway. Response to external stimulus suggests that the organism faces a condition atypical of the normal functioning of the tissues or organs, and it is directly linked to the occurrence and progression of numerous human diseases [19].

Deregulation of the PI3K‐Akt signalling pathway, cell migration and serine hydrolase activity has been reported as critical in cancer development [20, 21]. It has been established the role of serine hydrolases in multiple physiological processes in humans, including cancer [22], oxidative stress [23], metabolism [24] and bacterial infection [25]. Shields et al. [26] unravelled the retinoblastoma‐binding protein 9 (RBBP9) as a tumour‐associated serine hydrolase upregulated in pancreatic carcinomas patients. RBBP9 suppresses the activity of TGF‐β, which is implicated in carcinogenesis [27]. TGF‐β acts as a tumour suppressor by promoting apoptosis and inhibiting cell cycle progression [27].

We further investigated the interactions between the differentially expressed genes by constructing a PPI network. Attention was paid to the network module with the highest connectivity degree among all the resulting 78 modules from MCODE. This module contained 25 genes and 299 interactions, with a score of 24.92. GO and KEGG pathway analyses revealed the 25 genes to be enriched in metalloendopeptidase activity, metallopeptidase activity, hydrolase activity and Extracellular matrix organisation. Cheng et al. [28] underscored membrane metalloendopeptidase (MME) as an essential player in prostate cancer (PC) development. MME works in tandem with PTEN, a tumour suppressor, to control the activities of prostate progenitor cells and ultimately suppress PC progression [28]. Li et al. [29] also delineated the functions of MME in oesophageal squamous cell carcinoma progression. The roles of peptidases in tumourigenesis have been extensively studied by Young et al. [30] and Arrebola et al. [31].

The survival analyses of the 25 hub genes were performed using GEPIA2, and the genes that gave statistically significant results were ADAMTSL4, SEMA5B, ADAMTS15, THBS2, SPON1 and THBS1. F‐spondin 1(SPON1) forms a significant component of the extracellular matrix protein and promotes nerve precursor differentiation [32]. SPON1 also promotes the growth and development of axons in the peripheral nervous system and spinal cord [32]. Previous studies have shown that overexpression of SPON1 was linked to poor OS in bladder cancer patients [33], GC cohort [34] and in a PC cohort [35]. In hepatocellular carcinoma, microRNA‐506 binds to SPON1 to inhibit cell proliferation, migration and invasion [36]. SPON1 was found to be downregulated in colorectal cancers (CRC) compared to healthy colorectal tissues, which was associated with a better OS in CRC patients. SPON1 was found to be downregulated in DLBC and upregulated in Bladder Urothelial Carcinoma. However, the exact role of SPON1 in the development of BL remains to be explored.

The transmembrane protein Semaphorin‐5B (SEMA5B) belongs to a family of Semaphorins, which are known to play significant roles in axon growth and maturation during the development of the nervous system [37]. It regulates tumour growth and metastasis, as well as bone metastases and microvascular disorders [38]. SEMA5B has been demonstrated to contribute to renal cell carcinoma (RCC) progression [39]. The overexpression of SEMA5B has been linked to the development and proliferation of RCC cells and poor prognosis in gastric adenocarcinoma [40]. It has been proposed that Mitogen‐Activated Protein Kinases signalling, notch signalling and tumour signalling are the likely pathways regulated by SEMA5B in GC [40].

The thrombospondin 1 (THBS1) acts within the tumour microenvironment to stimulate tumour cell motility, regulates antitumour immunity, inhibits angiogenesis, control tumour growth factors and extracellular proteases [41]. Although THBS1 is not usually mutated in most cancers, its expression is regulated by multiple tumour suppressor genes and oncogenes, making it a major carcinogenic player [41]. The implication of THBS1 in human cancers is context‐dependent and has been reported in gastric carcinomas, bladder and CRC.

Zhang et al. [42] linked the overexpression of THBS1 to chemotherapy resistance and overall poor prognosis in patients with GC. Prior studies have demonstrated that overexpression of THBS1 and THBS2 leads to poor OS in GC patients and has the potential to act as both diagnostic and prognostic biomarkers for GC management [43]. These findings were consistent with Deng et al.'s [44] studies in another GC cohort.

Additionally, Berger et al. [45] reported THBS2 as a potential biomarker for managing pancreatic cancers. These studies provide evidence for the implication of THBSs in multiple human cancers. However, multiple independent studies have associated decreased expression of THBS1 with poor prognosis in several cancers, including GC [46], non‐small cell lung carcinoma [47] and oral squamous cell carcinomas [48]. Taken together, the expression of the THBSs in human cancers is context‐dependent and requires further research in other cancer types.

ADAMTSs (A Disintegrin and Metalloproteinase with Thrombospondin motifs) family of glycoproteins have been reported in multiple BP, including angiogenesis, cell migration, carcinogenesis, arthritis and coagulation [49]. Zhao et al. [50] showed ADAMTSL4 to be directly linked to poor prognosis in patients with Glioblastoma Multiforme and the expression of ADAMTSL4 to be correlated with immune responses such as infiltration of the tumour by immune cells. Deregulation of ADAMTSL4 has also been reported in nasopharyngeal carcinoma [51], acute lymphoblastic leukaemia [52] and oesophageal squamous cell carcinoma [53]. Low expression of ADAMTS15 was demonstrated to be associated with poor prognosis in breast carcinoma [54]. Moreover, studies by Viloria et al. [55] revealed that elevated expression of ADAMTS15 restricted the growth, invasion and metastasis of CRC cells. Binder et al. [56] also reported that the overexpression of ADAMTS15 in PC patients was directly linked with OS and consistent with findings from Molokwu et al. [57]. These independent studies collectively provide evidence that the expression levels of ADAMTSL4, SEMA5B, ADAMTS15, THBS2, SPON1 and THBS1 have the potential to serve as diagnostic and prognostic targets for the management of BL.

This study is limited by our inability to validate the biomarkers using molecular methods, such as quantitative polymerase chain reaction, Microarray and gene knockout analyses. We propose further research to be conducted to ascertain the functions of these makers in the progression of BL.

5. CONCLUSION

By analysing gene expression data from BL transcriptome sequencing projects, six biomarkers (ADAMTSL4, SEMA5B, ADAMTS15, THBS2, SPON1 and THBS1) were identified as possible indicators for diagnostic and prognostic targets for BL management. Moreover, these genes were implicated in other human cancers and significantly impacted patients' OS. Further research is warranted in mouse models to elucidate the exact roles of these genes in BL development and management.

AUTHOR CONTRIBUTIONS

Albert Doughan: Formal analysis; Validation; Investigation; Data Curation; Writing – Original Draft; Writing – Review & Editing. Samson Pandam Salifu: Conceptualisation; Validation; Resources; Writing – Review & Editing; Supervision; Funding acquisition.

CONFLICT OF INTEREST

The authors declare that they have no competing interests.

ACKNOWLEDGEMENTS

This work is supported by the US National Institutes of Health Common Fund (grant numbers U24HG006941 to Samson Pandam Salifu and 1U2RTW010679 to Samson Pandam Salifu and Albert Doughan). The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funding body did not play any roles in the design of the study, collection, analysis, and interpretation of data and in writing the manuscript.

Doughan, A. , Salifu, S.P. : Genes associated with diagnosis and prognosis of Burkitt lymphoma. IET Syst. Biol. 16(6), 220–229 (2022). 10.1049/syb2.12054

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in National Centre for Biotechnology Information Short Read Archive (NCBI‐SRA) under the Accession numbers SRP062178, SRP099346 and SRP009316.

REFERENCES

  • 1. Linch, D.C. : Burkitt lymphoma in adults. Br. J. Haematol. 156(6), 693–703 (2012). 10.1111/j.1365-2141.2011.08877.x [DOI] [PubMed] [Google Scholar]
  • 2. Dozzo, M. , et al.: Burkitt lymphoma in adolescents and young adults: management challenges. Adolesc. Health Med. Ther. 8, 11–29 (2017). 10.2147/ahmt.s94170 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Parkin, D.M. , et al.: Part I: cancer in Indigenous Africans—burden, distribution, and trends. Lancet Oncol. 9(7), 683–692 (2019). 10.1016/s1470-2045(08)70175-x [DOI] [PubMed] [Google Scholar]
  • 4. Grande, B.M. , et al.: Genome‐wide discovery of somatic coding and noncoding mutations in pediatric endemic and sporadic Burkitt lymphoma. Blood. 133(12), 1313–1324 (2019). 10.1182/blood-2018-09-871418 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Panea, R.I. , et al.: The whole‐genome landscape of Burkitt lymphoma subtypes. Blood J. Am. Soc. Hematol. 134(19), 1598–1607 (2019). 10.1182/blood.2019001880 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Kaymaz, Y. , et al.: Comprehensive transcriptome and mutational profiling of endemic Burkitt lymphoma reveals EBV type‐specific differences. Mol. Cancer Res. 15(5), 563–576 (2017). 10.1158/1541-7786.mcr-16-0305 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Abate, F. , et al.: Distinct viral and mutational spectrum of endemic Burkitt lymphoma. PLoS Pathog. 11(10), 1–21 (2015). 10.1371/journal.ppat.1005158 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Santiago, L. , et al.: Wnt signaling pathway protein LEF1 in cancer, as a biomarker for prognosis and a target for treatment. Am. J. Cancer Res. 7(6), 1389–1406 (2017) [PMC free article] [PubMed] [Google Scholar]
  • 9. Lombardo, K.A. , et al.: High‐throughput sequencing of the B‐cell receptor in African Burkitt lymphoma reveals clues to pathogenesis. Blood Adv. 1(9), 535–544 (2017). 10.1182/bloodadvances.2016000794 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Andrews, S. : FastQC: A Quality Control Tool for High Throughput Sequence Data (2010)
  • 11. Ewels, P. , et al.: MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 32(19), 3047–3048 (2016). 10.1093/bioinformatics/btw354 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Bolger, A.M. , Lohse, M. , Usadel, B. : Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 30(15), 2114–2120 (2014). 10.1093/bioinformatics/btu170 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Dobin, A. , et al.: STAR: ultrafast universal RNA‐seq aligner. Bioinformatics. 29(1), 15–21 (2013). 10.1093/bioinformatics/bts635 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Ge, S.X. , Jung, D. , Yao, R. : ShinyGO: a graphical gene‐set enrichment tool for animals and plants. Bioinformatics (Oxford, Engl and). 36(8), 2628–2629 (2020). 10.1093/bioinformatics/btz931 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Szklarczyk, D. , et al.: STRING v11: protein–protein association networks with increased coverage, supporting functional discovery in genome‐wide experimental datasets. Nucleic Acids Res. 47(D1), D607–D613 (2019). 10.1093/nar/gky1131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Shannon, P. , et al.: Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13(11), 2498–2504 (2003). 10.1101/gr.1239303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Bader, G.D. , Hogue, C.W. : An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinf. 4(1), 1–27 (2003) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Tang, Z. , et al.: GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 45(W1), W98–W102 (2017). 10.1093/nar/gkx247 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Lee, T.G. , et al.: Altered expression of cellular proliferation, apoptosis and the cell cycle‐related genes in lung cancer cells with acquired resistance to EGFR tyrosine kinase inhibitors. Oncol. Lett. 14(2), 2191–2197 (2017). 10.3892/ol.2017.6428 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Tan, X. , et al.: Tim‐4 promotes the growth of colorectal cancer by activating angiogenesis and recruiting tumor‐associated macrophages via the PI3K/AKT/mTOR signaling pathway. Oncol. Lett. 436, 119–128 (2018). 10.1016/j.canlet.2018.08.012 [DOI] [PubMed] [Google Scholar]
  • 21. Long, J.Z. , Cravatt, B.F. : The metabolic serine hydrolases and their functions in mammalian physiology and disease. Chem. Rev. 111(10), 6022–6063 (2011). 10.1021/cr200075y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Menendez, J.A. , Lupu, R. : Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nat. Rev. Cancer. 7(10), 763–777 (2007). 10.1038/nrc2222 [DOI] [PubMed] [Google Scholar]
  • 23. Kono, N. , et al.: Protection against oxidative stress‐induced hepatic injury by intracellular type II platelet‐activating factor acetylhydrolase by metabolism of oxidized phospholipids in vivo. J. Biol. Chem. 283(3), 1628–1636 (2008). 10.1074/jbc.m708622200 [DOI] [PubMed] [Google Scholar]
  • 24. Merkel, M. , Eckel, R.H. , Goldberg, I.J. : Lipoprotein lipase. J. Lipid Res. 43(12), 1997–2006 (2002). 10.1194/jlr.r200015-jlr200 [DOI] [PubMed] [Google Scholar]
  • 25. Lu, M. , et al.: Host inactivation of bacterial lipopolysaccharide prevents prolonged tolerance following gram‐negative bacterial infection. Cell Host Microbe. 4(3), 293–302 (2008). 10.1016/j.chom.2008.06.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Shields, D.J. , et al.: RBBP9: a tumor‐associated serine hydrolase activity required for pancreatic neoplasia. Proc. Natl. Acad. Sci. 107(5), 2189–2194 (2010). 10.1073/pnas.0911646107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Syed, V. : TGF‐β signaling in cancer. J. Cell. Biochem. 117(6), 1279–1287 (2016). 10.1002/jcb.25496 [DOI] [PubMed] [Google Scholar]
  • 28. Cheng, C.‐Y. , et al.: Membrane metalloendopeptidase suppresses prostate carcinogenesis by attenuating effects of gastrin‐releasing peptide on stem/progenitor cells. Oncogenesis. 9(3), 1–14 (2020). 10.1038/s41389-020-0222-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Li, M. , et al.: Membrane metalloendopeptidase (MME) suppresses metastasis of esophageal squamous cell carcinoma (ESCC) by inhibiting FAK‐RhoA signaling axis. Am. J. Pathol. 189(7), 1462–1472 (2019). 10.1016/j.ajpath.2019.04.007 [DOI] [PubMed] [Google Scholar]
  • 30. Young, M.‐J. , et al.: The role of ubiquitin‐specific peptidases in cancer progression. J. Biomed. Sci. 26(1), 1–14 (2019). 10.1186/s12929-019-0522-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Arrebola, Y.M. , et al.: Dipeptidyl peptidase IV and its implication in cancer. Biotecnol. Apl. 31(2), 93–101 (2014) [Google Scholar]
  • 32. Schubert, D. , et al.: F‐spondin promotes nerve precursor differentiation. J. Neurochem. 96(2), 444–453 (2006). 10.1111/j.1471-4159.2005.03563.x [DOI] [PubMed] [Google Scholar]
  • 33. Lv, J. , et al.: Mining TCGA database for tumor mutation burden and their clinical significance in bladder cancer. Biosci. Rep. 40(4), 1–12 (2020). 10.1042/bsr20194337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Ding, Y. , et al.: Identification of genes associated with gastric cancer survival and construction of a nomogram to improve risk stratification for patients with gastric cancer. Oncol. Lett. 20(1), 215–225 (2020). 10.3892/ol.2020.11543 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Wang, L.‐Y. , et al.: Biomarkers identified for prostate cancer patients through genome‐scale screening. Oncotarget. 8(54), 92055–92063 (2017). 10.18632/oncotarget.20739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Dai, W. , et al.: microRNA‐506 regulates proliferation, migration and invasion in hepatocellular carcinoma by targeting F‐spondin 1 (SPON1). Am. J. Cancer Res. 5(9), 2697–2707 (2015) [PMC free article] [PubMed] [Google Scholar]
  • 37. Yazdani, U. , Terman, J.R. : The semaphorins. Genome Biol. 7(3), 1–14 (2006). 10.1186/gb-2006-7-3-211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Jongbloets, B.C. , Pasterkamp, R.J. : Semaphorin signalling during development. Development. 141(17), 3292–3297 (2014). 10.1242/dev.105544 [DOI] [PubMed] [Google Scholar]
  • 39. Kundu, A. , et al.: PRDM16 suppresses HIF‐targeted gene expression in kidney cancer. J. Exp. Med. 217(6), 1–14 (2020). 10.1084/jem.20191005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Cao, H. , Song, X. , Hong, Y. : Increased expression of SEMA5B in gastric adenocarcinoma predicts poor prognosis: a study based on TCGA data. Chin. J. Oncol. 43(8), 856–860 (2021) [DOI] [PubMed] [Google Scholar]
  • 41. Isenberg, J.S. , Roberts, D.D. : THBS1 (thrombospondin‐1). Atlas Genet. Cytogenet. Oncol. 24(8), 291–299 (2020) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Zhang, X. , et al.: Upregulation of THBS1 is related to immunity and chemotherapy resistance in gastric cancer. Int. J. Gen. Med. 14, 4945–4957 (2021). 10.2147/ijgm.s329208 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Lu, Y. , et al.: Diagnostic, therapeutic, and prognostic value of the thrombospondin family in gastric cancer. Front. Mol. Biosci. 8, 146–154 (2021). 10.3389/fmolb.2021.647095 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Deng, S. , et al.: Selecting hub genes and predicting target genes of microRNAs in tuberculosis via the bioinformatics analysis. Genet. Res. 2021, 1–11 (2021). 10.1155/2021/6226291 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Berger, A.W. , et al.: A blood‐based multi marker assay supports the differential diagnosis of early‐stage pancreatic cancer. Theranostics. 9(5), 1280–1287 (2019). 10.7150/thno.29247 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Nakao, T. , et al.: Expression of thrombospondin‐1 and Ski are prognostic factors in advanced gastric cancer. Int. J. Clin. Oncol. 16(2), 145–152 (2011). 10.1007/s10147-010-0147-5 [DOI] [PubMed] [Google Scholar]
  • 47. Rouanne, M. , et al.: Osteopontin and thrombospondin‐1 play opposite roles in promoting tumor aggressiveness of primary resected non‐small cell lung cancer. BMC Cancer. 16(1), 1–13 (2016). 10.1186/s12885-016-2541-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Isenberg, J.S. , et al.: Regulation of nitric oxide signalling by thrombospondin 1: implications for anti‐angiogenic therapies. Nat. Rev. Cancer. 9(3), 182–194 (2009). 10.1038/nrc2561 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Le Goff, C. , Cormier‐Daire, V. : The ADAMTS (L) family and human genetic disorders. Hum. Mol. Genet. 20(R2), R163–R167 (2011). 10.1093/hmg/ddr361 [DOI] [PubMed] [Google Scholar]
  • 50. Zhao, Z. , et al.: ADAMTSL4, a secreted glycoprotein, is a novel immune‐related biomarker for primary glioblastoma multiforme. Dis. Markers. 2019, 1–12 (2019). 10.1155/2019/1802620 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Tao, Y. , et al.: Correlation of five secretory proteins with the nasopharyngeal carcinoma metastasis and the clinical applications. Oncotarget. 8(17), 29383–29394 (2017). 10.18632/oncotarget.14725 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52. Juric, D. , et al.: Differential gene expression patterns and interaction networks in BCR‐ABL–positive and–negative adult acute lymphoblastic leukemias. J. Clin. Oncol. 25(11), 1341–1349 (2007). 10.1200/jco.2006.09.3534 [DOI] [PubMed] [Google Scholar]
  • 53. Zhang, H. , et al.: An extracellular matrix‐based signature associated with immune microenvironment predicts the prognosis and therapeutic responses of patients with oesophageal squamous cell carcinoma. Front. Mol. Biosci. 8, 162 (2021). 10.3389/fmolb.2021.598427 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Porter, S. , et al.: ADAMTS8 and ADAMTS15 expression predicts survival in human breast carcinoma. Int. J. Cancer. 118(5), 1241–1247 (2006). 10.1002/ijc.21476 [DOI] [PubMed] [Google Scholar]
  • 55. Viloria, C.G. , et al.: Genetic inactivation of ADAMTS15 metalloprotease in human colorectal cancer. Cancer Res. 69(11), 4926–4934 (2009). 10.1158/0008-5472.can-08-4155 [DOI] [PubMed] [Google Scholar]
  • 56. Binder, M.J. , et al.: ADAMTS‐15 has a tumor suppressor role in prostate cancer. Biomolecules. 10(5), 682–891 (2020). 10.3390/biom10050682 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Molokwu, C.N. , et al.: Androgen regulates ADAMTS15 gene expression in prostate cancer cells. Cancer Invest. 28(7), 698–710 (2010). 10.3109/07357907.2010.489538 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are openly available in National Centre for Biotechnology Information Short Read Archive (NCBI‐SRA) under the Accession numbers SRP062178, SRP099346 and SRP009316.

Articles from IET Systems Biology are provided here courtesy of Wiley

RESOURCES