Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Exploring DNA methylation data for diagnostic classification of Diffuse large B-cell lymphoma in Dogs

Exploring DNA methylation data for diagnostic classification of Diffuse large B-cell lymphoma in... Diffuse large B-cell lymphoma (DLBCL) is a common B-lymphocyte tumor in dogs, making up 60-70% of cases. We assessed the utility of DNA methylation data for the diagnostic classification of DLBCL in dogs. We also assessed the utility of the classification features identified in cDLBCL for diagnostic classification of DLBCL in humans. The GSE94913 cDLBC L DNA methylation dataset from the Gene Expression Omnibus (GEO) was used for analysis. Differential methylation analysis was performed between the 37 cDLBCL and seven control lymph node samples in the dataset. 1701 differentially methylated probes were identified between the cDLBCL and control lymph nodes groups. Applying recursive feature elimination on the 1701 significant probes, 20 probes were selected for machine learning classification tasks. The methylation values of these 20 probes were used to build an SVM model and create the training and testing set. 100% of the test samples were accurately classified by the SVM model. The diagnostic classification utility of the identified differentially methylated CpGs/CDS was also assessed in humans using the GSE28094 human DLBCL dataset. 95% of 98 DLBCL and leukocyte samples obtained from this dataset was correctly classified using clustering techniques on 11 CpG sites of 5 genes (ERBB4, IGF2, PGF, PITX2, TJP1). The utility of DNA methylation data for the diagnostic classification of DLBCL in dogs is demonstrated. Further exploration of this data type for potential biomarker discovery in cDLBCL is necessary. Keywords: Biomarkers; Classification; DNA methylation; Dogs; Machine learning. © Giwa and Adu. This work is licensed under the Creative Commons Attribution-Non-Commercial-NoDerivs License 4.0 This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 8 Exploring DNA methylation data Giwa and Adu., 2023 cDLBCL and 7 control lymph node samples with the limma package in 1. Introduction R (version 4.1.1) (Ritchie et al., 2015). Significant probes were selected as those that met the threshold of p-adjusted value < 0.01 and logFC > Lymphoma is the most common hematological cancer in dogs (Mizuno et al., 2020). Its cause is unknown but genetic susceptibility and environmental factors are proposed to have contributory roles Machine learning (ML) (Zandvliet, 2016). Diffuse large B-cell lymphocytes (DLBCL) is the Clustering of all 44 samples based on all significant probes was done most common tumor of B-lymphocytes in dogs, making up about 60- using the factoextra R package (version 1.0.7) (Kassambara and 70% of all cases (Ferraresso et al., 2017). The advances in treatment Mundt, 2020). A combination of k-means and hierarchical clustering options have been modest, with high relapse rates and a low 2-year algorithm was applied to the significant probes (Kassambara and survival rate of 20% (Aresu, 2019; Martini et al., 2019). Novel Mundt, 2020). Thereafter, the resulting data object was visualized as a therapeutic approaches is therefore needed. Canine DLBCL (cDLBCL) dendogram (Kassambara and Mundt, 2020). shares similar molecular and clinical features with human DLBCL (hDLBCL) and is therefore a comparative model (Richards et al., 2013; To find possible features diagnostic of cDLBCL, machine learning Aresu et al., 2019). Knowledge obtained from studying cDLBCL could classification tasks were performed. Recursive feature elimination be beneficial in understanding hDLBCL disease. Clinically cDLBCL (RFE) was performed on the significant differentially methylated presents as a moderate to severe peripheral lymph node enlargement probes to obtain the top 20 optimal features for classification. ML (Aresu et al., 2015). Diagnosis is made using a combination of flow model creation and classification was done using LibSVM (Chang and cytometry, cytology and histopathology (Riondato and Comazzi, Lin, 2011). The feature values were scaled with a LibSVM built-in 2021). Prognosis and selection of optimal treatment methods after python script. Using the train_test_split function in Scikit-learn diagnosis is still a challenge (Klimiuk et al., 2021). (Pedregosa et al., 2011), the data was splitted into training (70%) and testing (30%) sets. The SVM algorithm was used for the training and Molecular investigation at different omics levels have shown cDLBCL evaluation of the model. To determine the best parameters of the SVM to be a heterogeneous tumor (Richards et al., 2013; Giannuzzi et al., model to be subsequently applied to the test samples, 5-fold cross- 2019; Sirivisoot et al., 2022). Tumor and non-tumor samples were validation was performed. The model was applied to predict the class separated into different clusters, by the application of clustering of the samples in the testing set. The model was evaluated using methods on gene expression data (Mudaliar et al., 2013). Gene precision, recall and accuracy. Also, interaction analysis of the twenty Expression profiling and immunohistochemistry was used to genes identified above was performed with STRING database (version distinguish the subtypes of hDLBCL and cDLBCL (Richards et al., 11.5) (Szklarczyk et al., 2021). Using TRRUST (version 2) (Han et al., 2013). Epigenetic and genetic alterations are necessary for tumor 2018), we performed a transcriptional regulatory analysis to find the development (Takeshima and Ushijima, 2019). Alterations in the transcription factors regulating the identified genes. epigenome during tumorigenesis include global changes in histone modification marks (Fraga et al., 2005; Hosseini and Minucci, 2017), To assess whether the methylated probes identified in this study could deregulation of noncoding RNA networks (Liz and Esteller, 2016), be useful for diagnostic classification of hDLBCL, RFE was done to global DNA methylation loss and CpG promoter islands select the top 200 important features for classification. This was to hypermethylation of tumor suppressor genes (Baylin and Jones, accommodate for the highly reduced gene number in the human 2016). Few studies have been done to characterize the epigenetic dataset to be used for this assessment. The GSE28094 dataset was landscape of cDLBCL and these have been focused on DNA retrieved from the GEO database (Fernandez et al., 2012). This dataset methylation. These included DNA methylation studies that focused on is a human DNA methylation dataset comprising 424 normal tissue single gene (Sato et al., 2014; Fujiwara-Igarashi et al., 2014; Tomiyasu samples, 1054 tumor samples and 150 non-cancerous disorders et al., 2014; Ferraresso et al., 2014), and genome-wide (Ferraresso et samples. The beta values of 1505 CpG sites corresponding to 808 al., 2017; Aresu et al., 2019; Hsu et al., 2021). The full understanding of genes were profiled in the dataset. We extracted the profiles for 49 the molecular mechanisms underlying cDLBCL is still incomplete. This DLBCL and 49 healthy normal leukocytes from the dataset. The study aims to explore the utility of DNA methylation data for available genes out of the 200 top genes identified from RFE were diagnostic classification of DLBCL in dogs. The identified classification thereafter used as classification features. Clustering of these 98 human features would also be assessed for diagnostic classification of DLBCL samples was then performed using factoextra package (Kassambara in humans. and Mundt, 2020). Interaction analysis and transcriptional regulatory analysis of the genes were performed using STRING database (version 2. Experimental 11.5) (Szklarczyk et al., 2021) and TRRUST (version 2) (Han et al., 2018) respectively. Python and R scripts used in this study’s analyses Dataset are available in https://github.com/ZEB-LASU/cDLBCL The GSE94913 cDLBCL DNA methylation dataset was retrieved from the GEO database (Ferraresso et al., 2017). This dataset is composed of 3. Results 37 cDLBCL samples and 7 control lymph node samples. This dataset is The differential methylation analysis between the cDLBCL and control a microarray of 37479 CpG and coding sequences (CDS) probes, with lymph node groups identified 1701 differentially methylated probes. standard microarray data normalization and processing applied. 100% of the 44 samples were accurately clustered using these 1701 significantly methylated probes (fig. 1). The top 20 probes for Differential methylation analysis classifying samples selected by RFE are shown in Table 1. Annotation Differential methylation analysis was performed between the 37 9| This journal is © The Nigerian Young Academy 2022 A nn a ls of Science and Technology 2023 Vol. 8 (1): 8-15 Exploring DNA methylation data Giwa and Adu., 2023 on the Ensembl genome browser (CanFam 3.1) identified these to To assess the diagnostic classification utility of the methylated genes include eleven protein-coding genes, three non coding RNAs, and six in hDLBCL, 11 CpG sites of six genes (ERBB4, IGF2, PGF, PITX2, TJP1, unidentified genes (Table 1). These were used in ML training and test PDGFA) was used as features for clustering the 98 samples obtained set building. The training set was built on 30 samples randomly from the GSE28094 dataset. 95% of the samples were accurately selected by the train_test_split Scikit-learn function, comprising of 25 clustered into their respective groups (fig. 2). The 5 misclassified cDLBCL samples and 5 control lymph node samples. The testing set samples were however in a separate cluster group even though placed was made up of 14 samples selected by train_test_split Scikit-learn in the same branch with the normal leukocyte tissues (fig. 2). A function, comprising of 12 cDLBCL samples and 2 control lymph node supervised machine learning model could not be constructed due to samples. The 5-fold cross-validation accuracy obtained was 100%. the difference in platforms and the small number of genes profiled in Evaluation of the SVM model on the testing set resulted in an accuracy the available hDLBCL dataset. The interaction analysis between the six of 100% of correctly predicted samples (Table 2). For the interaction genes identified interactions between four of the six genes (PDGFA, and transcriptional regulatory analysis, no interaction and PGF, ERBB4, IGF2) (fig. 3). The transcriptional regulatory analysis transcriptional regulatory relationships were returned. revealed WT1 regulating the genes IGF2 and PDGFA. Fig. 1. Clustering of the cDLBCL and control lymph node samples using all 1701 significant methylation probes. The x-axis contains the samples and which are labelled for clear identification. Identified were two principal clusters with the 7 control lymph node samples all grouped in the left cluster (black rectangle box) while the 37 cDLBCL samples were all grouped in the right cluster (red rectangle box). DNA methylation data for diagnostic classification of DLBCL in 4. Discussion Applying machine learning classification techniques to cDLBCL DNA methylation data can give good results and thus can be beneficial The differential methylation analysis revealed expectedly that there Dysregulated expression of some of these identified top genes have were differences in methylation between cDLBCL and the control been associated with several cancers including RAB7A in gastric samples. Indeed, aberrant DNA methylation is characteristic of cancer cancer (Liu et al., 2020), RAB7A in breast cancer (Xie et al., 2019), (Dong et al., 2014), and cDLBCL (Ferraresso et al., 2017; Fujiwara- LEF1 in melanoma (Kim et al., 2020), IGF2 in hepatocellular Igarashi et al., 2014; Hsu et al., 2021; Sato et al., 2018). The accuracy of carcinoma (Ma et al., 2019) and adrenocortical carcinoma (Pereira et al., 2019), ST8SIA2 in lung cancer (Hao et al., 2019), DDIT4 in acute the clustering despite using all 1701 significant probes substantiated myeloid leukemia (Cheng et al., 2020) and in ovarian carcinoma the differential methylation analysis result and indicated their (Chang et al., 2018), and PPM1H in colon adenocarcinoma (Sugiura et possible utility for classification. The high accuracy obtained by the ML al., 2008) and in pancreatic cancer (Zhu et al., 2016). These genes model (Table 2) constructed using 20 significant probes (features) could be potential drivers of cDLBCL in Dogs. Aberrant methylation of these genes could affect their expression. They can also serve as demonstrate the validity of these CpGs for diagnostic identification of possible therapeutic targets upon further studies and functional DLBCL in dogs. As a limitation, the testing set was small, requiring studies of these identified genes should elucidate our understanding validation in larger sample cohorts. We were able to explore the use of DNA methylation data for diagnostic classification of DLBCL in Dogs. This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 10 Exploring DNA methylation data Giwa and Adu., 2023 Table 1. Twenty probes selected by recursive feature elimination and used as features for machine learning classification Probe ID adj-p-value Gene name GT_chr6:31498360-31499687_296737 2.40686e-05 TNP2* GT_CpG_102:chr20:2792244-2793806_10609 5.00947e-19 RAB7A* GT_CpG_133:chr8:34699507-34700781_125659 4.57457e-07 JKAMP* GT_CpG_199:chr32:28632693-28635175_287611 1.84638e-06 LEF1* GT_CpG_208:chr10:6070361-6072996_308205 2.76684e-22 PPM1H* GT_CpG_217:chr3:47653366-47655868_330479 4.13005e-17 ST8SIA2* GT_CpG_31:chr4:22936675-22937328_508162 6.84959e-12 DDIT4* GT_CpG_35:chr25:15429047-15429606_825639 7.52139e-17 SGCG* GT_CpG_42:chr18:46302185-46302823_622168 5.85738e-12 IGF2* GT_CpG_57:chr22:8708474-8709185_748846 3.30485e-08 DGKH* GT_CpG_62:chr6:17872282-17873207_782319 1.75655e-12 ENSCAFT00000077585** GT_CpG_72:chr10:55372722-55373751_851800 1.48866e-14 ENSCAFT00000065559** GT_CpG_82:chr19:23852100-23852984_905754 0.00054 ENSCAFT00000089251** GT_CpG_90:chr1:104706255-104707295_948211 3.10691e-19 ENSCAFT00000044531* GT_CpG_103:chr9:5445423-5446725_17310 8.35098e-18 GT_CpG_33:chr23:33251449-33252124_530445 2.76684e-22 GT_CpG_43:chr15:45025176-45025844_631179 6.43581e-19 GT_CpG_69:chr11:62398353-62399319_830114 1.94461e-17 GT_CpG_77:chr27:31835085-31836199_881657 1.55622e-14 GT_CpG_99:chr6:38018350-38019848_992235 1.37342e-07 Shown are the probe identifiers, their adjusted p-values, and corresponding gene names. *protein coding genes. **long non-coding RNA gene. Blanks are genes without annotation names Table 2: Machine learning classification results of the testing set using SVM model Class Precision Recall cDLBCL 1.0 1.0 Control lymph node 1.0 1.0 Accuracy 100% (14/14) 11| This journal is © The Nigerian Young Academy 2022 A nn als of Science and Technology 2023 Vol. 8 (1): 8-15 Exploring DNA methylation data Giwa and Adu., 2023 Fig. 2. Clustering of the 98 DLBCL and leukocyte samples of the GSE28094 human dataset. The x-axis contains the samples, with the DLBCL samples labelled with D and leukocyte samples labelled with HL. Identified were two main clusters with the DLBCL samples predominantly clustered in the right cluster (red rectangle box) while the leukocyte samples were predominantly clustered in the left cluster (black rectangle box). 5 DLBCL samples were misclassified. Fig. 3. Gene interaction analysis of the six (ERBB4, IGF2, PGF, PITX2, TJP1, PDGFA) genes used for clustering the hDLBCL samples. understanding of DLBCL disease mechanisms in dogs. The be potentially used for diagnostic identification of DLBCL in humans unavailability of external datasets to further validate this CpG upon further validation. WT1 was reported to have transcriptionally signature indicates the need for more investigative studies to regulate IGF2 and PDGFA, and has been associated with several characterize the epigenetic landscape of DLBCL in dogs. The cancers (Qi et al., 2015; Zhang et al., 2020). Diagnostic classification of unsupervised model based on 11 CpGs of six genes (ERBB4, IGF2, PGF, tumors using DNA methylation data has been demonstrated in PITX2, TJP1, PDGFA) was able to discriminate hDLBCL samples and multiple human cancers (Capper et al., 2018; Liu et al., 2020; Giwa et leukocyte samples with high accuracy (fig 2). This reinforces the al., 2021). It could therefore be possible to develop diagnostic molecular and clinical similarity between cDLBCL and hDLBCL biomarkers that are clinically useful in canine and human DLBCL. (Richards et al., 2013; Aresu et al., 2019). These six genes could thus This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 12 Giwa and Adu., 2023 Exploring DNA methylation data 5. Conclusion Chen, Y., Hesla, A.C., Lin, Y., Ghaderi, M., Liu, M., Yang, C., Zhang, Y., This exploratory study demonstrates the utility of DNA methylation Tsagkozis, P., Larsson, O., and Haglund, F., (2020), Transcriptome data for diagnostic classification of DLBCL in dogs. DNA methylation profiling of Ewing sarcomas - treatment resistance pathways and IGF- data-based diagnostics for tumor classification is becoming dependency. Molecular Oncology, 14(5):1101-1117. mainstream. Therefore, characterizing the methylation landscape of DLBCL in dogs should have clinical benefit. It will also further Cheng, Z., Dai, Y., Pang, Y., Jiao, Y., Liu, Y., Cui, L., Quan, L., Qian, T., Zeng, understanding of the disease mechanisms and aid therapeutic T., Si, C., Huang, W., Chen, J., Pang, Y., Ye, X., Shi, J., and Fu, L., (2020), development as survival rate of DLBCL in dogs is still low. Further Up-regulation of DDIT4 predicts poor prognosis in acute myeloid exploration of this data type for potential biomarker discovery in leukaemia. Journal of Cellular and Molecular Medicine, 24(1):1067- cDLBCL is necessary. Declaration of Conflict of Interests Dong, Y., Zhao, H., Li, H., Li, X., and Yang, S., (2014), DNA methylation as No conflict of interests to declare an early diagnostic marker of cancer (Review). Biomedical Reports, 2(3):326-330. Authors’ Contributions Fernandez, A.F., Assenov, Y., Martin-Subero, J.I., Balint, B., Siebert, R., Conception: [AG] Taniguchi, H., Yamamoto, H., Hidalgo, M., Tan, A.C., Galm, O., Ferrer, I., Design: [AG] Sanchez-Cespedes, M., Villanueva, A., Carmona, J., Sanchez-Mut, J.V., Execution: [OA, AG] Berdasco, M., Moreno, V., Capella, G., Monk, D., Ballestar, E., Ropero, S., Interpretation: [OA, AG] Martinez, R., Sanchez-Carbayo, M., Prosper, F., Agirre, X., Fraga, M.F., Writing the paper: [OA, AG] Graña, O., Perez-Jurado, L., Mora, J., Puig, S., Prat, J., Badimon, L., Puca, A.A., Meltzer, S.J., Lengauer, T., Bridgewater, J., Bock, C., and Esteller, M., (2012), A DNA methylation fingerprint of 1628 human samples. References Genome Research, 22(2):407-419. Aresu, L., Martini, V., Rossi, F., Vignoli, M., Sampaolo, M., Aricò, A., Ferraresso, S., Bresolin, S., Aricò, A., Comazzi, S., Gelain, M.E., Riondato, Laganga, P., Pierini, A., Frayssinet, P., Mantovani, R., and Marconato, L., F., Bargelloni, L., Marconato, L., Kronnie, Gt., and Aresu, L., (2014), (2015), Canine indolent and aggressive lymphoma: clinical spectrum Epigenetic silencing of TFPI-2 in canine diffuse large B-cell lymphoma. with histologic correlation. Veterinary and Comparative Oncology, PLoS One, 9(4):e92707. 13(4):348-362. Ferraresso, S., Aricò, A., Sanavia, T., Da Ros, S., Milan, M., Cascione, L., Aresu, L., (2016), Canine Lymphoma, More than a morphological Comazzi, S., Martini, V., Giantin, M., Di Camillo, B., Mazzariol, S., diagnosis: what we have learned about diffuse large B-cell lymphoma. Giannuzzi, D., Marconato, L., and Aresu, L., (2017), DNA methylation Frontiers in Veterinary Science, 3:77. profiling reveals common signatures of tumorigenesis and defines epigenetic prognostic subtypes of canine Diffuse Large B-cell Aresu, L., Ferraresso, S., Marconato, L., Cascione, L., Napoli, S., Gaudio, Lymphoma. Scientific Reports, 7(1):11591. E., Kwee, I., Tarantelli, C., Testa, A., Maniaci, C., Ciulli, A., Hillmann, P., Bohnacker, T., Wymann, M.P., Comazzi, S., Milan, M., Riondato, F., Fraga, M.F., Ballestar, E., Villar-Garea, A., Boix-Chornet, M., Espada, J., Rovere, G.D., Giantin, M., Giannuzzi D., and Bertoni, F., (2019), New Schotta, G., Bonaldi, T., Haydon, C., Ropero, S., Petrie, K., Iyer, N.G., molecular and therapeutic insights into canine diffuse large B-cell Pérez-Rosado, A., Calvo, E., Lopez, J.A., Cano, A., Calasanz, M.J., Colomer, lymphoma elucidates the role of the dog as a model for human disease. D., Piris, M.A., Ahn, N., Imhof, A., Caldas, C., Jenuwein, T., and Esteller, Haematologica, 104(6):e256-e259. M., (2005), Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human cancer. Nature Genetics, Baylin, S.B., and Jones, P.A., (2016), Epigenetic determinants of cancer. 37(4):391-400. Cold Spring Harbor Perspectives in Biology, 8(9):a019505. Fujiwara-Igarashi, A., Goto-Koshino, Y., Mochizuki, H., Sato, M., Fujino, Capper, D., Jones, D.T.W., Sill, M., Hovestadt, V., Schrimpf, D., et al. Y., Ohno, K., and Tsujimoto, H., (2014), Inhibition of p16 tumor (2018), DNA methylation-based classification of central nervous suppressor gene expression via promoter hypermethylation in canine system tumours. Nature, 555(7697):469-474. lymphoid tumor cells. Research in Veterinary Science, 97(1):60-63. Chang, B., Meng, J., Zhu, H., Du, X., Sun, L., Wang, L., Li, S., and Yang, G., Giannuzzi, D., Marconato, L., Cascione, L., Comazzi, S., Elgendy, R., (2018), Overexpression of the recently identified oncogene REDD1 Pegolo, S., Cecchinato, A., Bertoni, F., Aresu, L., and Ferraresso, S., correlates with tumor progression and is an independent unfavorable (2019), Mutational landscape of canine B-cell lymphoma profiled at prognostic factor for ovarian carcinoma. Diagnostic Pathology, single nucleotide resolution by RNA-seq. PLoS ONE, 14(4): e0215154. 13(1):87. Chang, C.C., and Lin, C.J., (2011), LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:27 13| This journal is © The Nigerian Young Academy 2022 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 Giwa and Adu., 2023 Exploring DNA methylation data Giwa, A., Rossouw, S.C., Fatai, A., Gamieldien, J., Christoffels, A., and Martini, V., Aresu, L., Riondato, F., Marconato, L., Cozzi, M., Stefanello, Bendou, H., (2021), Predicting amplification of MYCN using CpG D., and Comazzi, S., (2019), Prognostic role of non-neoplastic methylation biomarkers in neuroblastoma. Future Oncology, lymphocytes in lymph node aspirates from dogs with diffuse large B- 17(34):4769-4783. cell lymphoma treated with chemo-immunotherapy. Research in Veterinary Science, 125:130-135. Han, H., Cho, J.W., Lee, S., Yun, A., Kim, H., Bae, D., Yang, S., Kim, C.Y., Lee, M., Kim, E., Lee, S., Kang, B., Jeong, D., Kim, Y., Jeon, H.N., Jung, H., Mizuno, T., Kato, Y., Kaneko, M.K., Sakai, Y., Shiga, T., Kato, M., Tsukui, Nam, S., Chung, M., Kim, J.H., and Lee, I., (2018), TRRUST v2: an T., Takemoto, H., Tokimasa, A., Baba, K., Nemoto, Y., Sakai, O., and expanded reference database of human and mouse transcriptional Igase, M., (2020), Generation of a canine anti-canine CD20 antibody for regulatory interactions. Nucleic Acids Research, 46(D1):D380-D386. canine lymphoma treatment. Scientific Reports, 10(1):11476. Hao, J., Zeltz, C., Pintilie, M., Li, Q., Sakashita, S., Wang, T., Cabanero, M., Mudaliar, M.A., Haggart, R.D., Miele, G., Sellar, G., Tan, K.A., Goodlad, Martins-Filho, S.N., Wang, D.Y., Pasko, E., Venkat, K., Joseph, J., J.R., Milne, E., Vail, D.M., Kurzman, I., Crowther, D., and Argyle, D.J., Raghavan, V., Zhu, C.Q., Wang, Y.H., Moghal, N., Tsao, M.S., and Navab, (2013), Comparative gene expression profiling identifies common R., (2019), Characterization of distinct populations of carcinoma- molecular signatures of NF-κB activation in canine and human diffuse associated fibroblasts from non-small cell lung carcinoma reveals a large B cell lymphoma (DLBCL). PLoS One, 8(9):e72591. role for ST8SIA2 in cancer cell invasion. Neoplasia, 21(5):482-493. Pedregosa F, Varoquaux G, Gramfort A, Michel V., Thirion, B., Grisel, O., Hosseini, A., and Minucci, S., (2017), A comprehensive review of Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., lysine-specific demethylase 1 and its roles in cancer. Epigenomics, Passos, A., Cournapeau, D., Brucher, M., Perrot, M., and Duchesnay, E., 9(8):1123-1142. (2011), Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12:2825–2830. Hsu, C.H., Tomiyasu, H., Lee, J.J., Tung, C.W., Liao, C.H., Chuang, C.H., Huang, L.Y., Liao, K.W., Chou, C.H., Liao, A.T.C., and Lin, C.S., (2021), Pereira, S.S., Monteiro, M.P., Costa, M.M., Moreira, Â., Alves, M.G., Genome-wide DNA methylation analysis using MethylCap-seq in Oliveira, P.F., Jarak. I., and Pignatelli, D., (2019), IGF2 role in canine high-grade B-cell lymphoma. Journal of Leukocyte Biology, adrenocortical carcinoma biology. Endocrine, 66(2):326-337. 109(6):1089-1103. Qi, X.W., Zhang, F., Wu, H., Liu, J.L., Zong, B.G., Xu, C., Jiang, J., (2015), Kassambara, A., and Mundt, F., (2020), Factoextra R package. Wilms' tumor 1 (WT1) expression and prognosis in solid cancer https://CRAN.R-project.org/package=factoextra patients: a systematic review and meta-analysis. Scientific Reports, 5:8924. Kim, G.H., Fang, X.Q., Lim, W.J., Park, J., Kang, T.B., Kim, J.H., and Lim, J.H., (2020), Cinobufagin suppresses melanoma cell growth by Richards, K.L., Motsinger-Reif, A.A., Chen, H.W., Fedoriw, Y., Fan, C., inhibiting LEF1. International Journal of Molecular Sciences, Nielsen, D.M., Small, G.W., Thomas, R., Smith, C., Dave, S.S., Perou, C.M., 21(18):6706. Breen, M., Borst, L.B., and Suter, S.E., (2013), Gene profiling of canine B-cell lymphoma reveals germinal center and postgerminal center Klimiuk, P., Łopuszynski, W., Bulak, K., and Brzana, A., (2021), subtypes with different survival times, modeling human DLBCL. Evaluation of the proliferative activity of diffuse large b-cell lymphoma Cancer Research, 73(16):5029-5039. (DLBCL) in dogs with respect to patient eligibility for anthracycline- based chemotherapy. Animals (Basel), 11(4):1183. Riondato, F., and Comazzi, S., (2021), Flow Cytometry in the Diagnosis of Canine B-Cell Lymphoma. Frontiers in Veterinary Science, Liu, H., Xu, J., Yao, Q., Zhang, Z., Guo, Q., and Lin, J., (2020), Rab7 is 8:600986. associated with poor prognosis of gastric cancer and promotes proliferation, invasion, and migration of gastric cancer cells. Medical Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, Science Monitor, 26:e922217. G.K., (2015), limma powers differential expression analyses for RNA- sequencing and microarray studies. Nucleic Acids Research, 43(7):e47. Liu, M.C., Oxnard, G.R., Klein, E.A., Swanton, C., Seiden, M.V.; CCGA Consortium, (2020), Sensitive and specific multi-cancer detection and Sato, M., Mochizuki, H., Goto-Koshino, Y., Fujiwara-Igarashi, A., localization using methylation signatures in cell-free DNA. Annals of Takahashi, M., Fujino, Y., Ohno, K., and Tsujimoto, H., (2014), Oncology, 31(6):745-759. Hypermethylation of the death-associated protein kinase CpG island in canine B-cell lymphoid tumors. Veterinary Immunology and Liz, J., and Esteller, M., (2016), lncRNAs and microRNAs with a role in Immunopathology, 161(3-4):222-231. cancer development. Biochimica et Biophysica Acta, 1859(1):169–176. Sato, M., Mochizuki, H., Goto-Koshino, Y., Fujiwara-Igarashi, A., Ma, Y., Chen, Y., Chen, L., Liu, Z., Ieong, M.L., Gao, F., and Huang, W., Takahashi, M., Ohno, K., and Tsujimoto, H., (2018), Prognostic (2018), Promotion of Insulin-like growth factor II in cell proliferation significance of hypermethylation of death-associated protein kinase and epithelial-mesenchymal transition in hepatocellular carcinoma. (DAPK) gene CpG island in dogs with high-grade B-cell lymphoma. Journal of Cancer Research and Therapeutics, 14(4):844-850. Veterinary and Comparative Oncology, 16(3):409-415. Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 14 This journal is © The Nigerian Young Academy 2023 Exploring DNA methylation data Giwa and Adu., 2023 Sirivisoot, S., Kasantikul, T., Techangamsuwan, S., Radtanakatikanon, Tomiyasu, H., Fujiwara-Igarashi, A., Goto-Koshino, Y., Fujino, Y., Ohno, A., Chen, K., Lin, T.Y., and Rungsipipat, A., (2022), Evaluation of 41 K., and Tsujimoto, H., (2014), Evaluation of DNA methylation profiles single nucleotide polymorphisms in canine diffuse large B-cell of the CpG island of the ABCB1 gene in dogs with lymphoma. American lymphomas using MassARRAY. Scientific Reports, 12(1):5120. Journal of Veterinary Research, 75(9):835-841. Sugiura, T., Noguchi, Y., Sakurai, K., and Hattori, C., (2008), Protein Xie, J., Yan, Y., Liu, F., Kang, H., Xu, F., Xiao, W., Wang H., and Wang, Y., phosphatase 1H, overexpressed in colon adenocarcinoma, is (2019), Knockdown of Rab7a suppresses the proliferation, migration, associated with CSE1L. Cancer Biology and Therapy, 7(2):285-292. and xenograft tumor growth of breast cancer cells. Bioscience Reports, 39(2):BSR20180480. Szklarczyk, D., Gable, A.L., Nastou, K.C., Lyon, D., Kirsch, R., Pyysalo, S., Doncheva, N.T., Legeay, M., Fang, T., Bork, P., Jensen, L.J., and von Zandvliet, M., (2016), Canine lymphoma: a review. Veterinary Mering, C., (2021), The STRING database in 2021: customizable Quarterly, 36(2):76-104. protein-protein networks, and functional characterization of user- uploaded gene/measurement sets. Nucleic Acids Research, Zhang, Y., Yan, W.T., Yang, Z.Y., Li, Y.L., Tan, X.N., Jiang, J., Zhang, Y., and 49(D1):D605-D612. Qi, X.W., (2020), The role of WT1 in breast cancer: clinical implications, biological effects and molecular mechanism. Takeshima, H., and Ushijima, T., (2019), Accumulation of genetic and International Journal of Biological Sciences, 16(8):1474-1480. epigenetic alterations in normal cells and cancer risk. NPJ Precision Oncology, 3:7. Zhu, H., Qin, H., Li, D.M., Liu, J., and Zhao, Q., (2016), Effect of PPM1H on malignant phenotype of human pancreatic cancer cells. Oncology Reports, 36(5):2926-2934. 15| This journal is © The Nigerian Young Academy 2022 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Science and Technology de Gruyter

Exploring DNA methylation data for diagnostic classification of Diffuse large B-cell lymphoma in Dogs

Loading next page...
 
/lp/de-gruyter/exploring-dna-methylation-data-for-diagnostic-classification-of-0i1OwHhxWe

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
de Gruyter
Copyright
© 2023 Abdulazeez Giwa et al., published by Sciendo
eISSN
2544-6320
DOI
10.2478/ast-2023-0002
Publisher site
See Article on Publisher Site

Abstract

Diffuse large B-cell lymphoma (DLBCL) is a common B-lymphocyte tumor in dogs, making up 60-70% of cases. We assessed the utility of DNA methylation data for the diagnostic classification of DLBCL in dogs. We also assessed the utility of the classification features identified in cDLBCL for diagnostic classification of DLBCL in humans. The GSE94913 cDLBC L DNA methylation dataset from the Gene Expression Omnibus (GEO) was used for analysis. Differential methylation analysis was performed between the 37 cDLBCL and seven control lymph node samples in the dataset. 1701 differentially methylated probes were identified between the cDLBCL and control lymph nodes groups. Applying recursive feature elimination on the 1701 significant probes, 20 probes were selected for machine learning classification tasks. The methylation values of these 20 probes were used to build an SVM model and create the training and testing set. 100% of the test samples were accurately classified by the SVM model. The diagnostic classification utility of the identified differentially methylated CpGs/CDS was also assessed in humans using the GSE28094 human DLBCL dataset. 95% of 98 DLBCL and leukocyte samples obtained from this dataset was correctly classified using clustering techniques on 11 CpG sites of 5 genes (ERBB4, IGF2, PGF, PITX2, TJP1). The utility of DNA methylation data for the diagnostic classification of DLBCL in dogs is demonstrated. Further exploration of this data type for potential biomarker discovery in cDLBCL is necessary. Keywords: Biomarkers; Classification; DNA methylation; Dogs; Machine learning. © Giwa and Adu. This work is licensed under the Creative Commons Attribution-Non-Commercial-NoDerivs License 4.0 This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 8 Exploring DNA methylation data Giwa and Adu., 2023 cDLBCL and 7 control lymph node samples with the limma package in 1. Introduction R (version 4.1.1) (Ritchie et al., 2015). Significant probes were selected as those that met the threshold of p-adjusted value < 0.01 and logFC > Lymphoma is the most common hematological cancer in dogs (Mizuno et al., 2020). Its cause is unknown but genetic susceptibility and environmental factors are proposed to have contributory roles Machine learning (ML) (Zandvliet, 2016). Diffuse large B-cell lymphocytes (DLBCL) is the Clustering of all 44 samples based on all significant probes was done most common tumor of B-lymphocytes in dogs, making up about 60- using the factoextra R package (version 1.0.7) (Kassambara and 70% of all cases (Ferraresso et al., 2017). The advances in treatment Mundt, 2020). A combination of k-means and hierarchical clustering options have been modest, with high relapse rates and a low 2-year algorithm was applied to the significant probes (Kassambara and survival rate of 20% (Aresu, 2019; Martini et al., 2019). Novel Mundt, 2020). Thereafter, the resulting data object was visualized as a therapeutic approaches is therefore needed. Canine DLBCL (cDLBCL) dendogram (Kassambara and Mundt, 2020). shares similar molecular and clinical features with human DLBCL (hDLBCL) and is therefore a comparative model (Richards et al., 2013; To find possible features diagnostic of cDLBCL, machine learning Aresu et al., 2019). Knowledge obtained from studying cDLBCL could classification tasks were performed. Recursive feature elimination be beneficial in understanding hDLBCL disease. Clinically cDLBCL (RFE) was performed on the significant differentially methylated presents as a moderate to severe peripheral lymph node enlargement probes to obtain the top 20 optimal features for classification. ML (Aresu et al., 2015). Diagnosis is made using a combination of flow model creation and classification was done using LibSVM (Chang and cytometry, cytology and histopathology (Riondato and Comazzi, Lin, 2011). The feature values were scaled with a LibSVM built-in 2021). Prognosis and selection of optimal treatment methods after python script. Using the train_test_split function in Scikit-learn diagnosis is still a challenge (Klimiuk et al., 2021). (Pedregosa et al., 2011), the data was splitted into training (70%) and testing (30%) sets. The SVM algorithm was used for the training and Molecular investigation at different omics levels have shown cDLBCL evaluation of the model. To determine the best parameters of the SVM to be a heterogeneous tumor (Richards et al., 2013; Giannuzzi et al., model to be subsequently applied to the test samples, 5-fold cross- 2019; Sirivisoot et al., 2022). Tumor and non-tumor samples were validation was performed. The model was applied to predict the class separated into different clusters, by the application of clustering of the samples in the testing set. The model was evaluated using methods on gene expression data (Mudaliar et al., 2013). Gene precision, recall and accuracy. Also, interaction analysis of the twenty Expression profiling and immunohistochemistry was used to genes identified above was performed with STRING database (version distinguish the subtypes of hDLBCL and cDLBCL (Richards et al., 11.5) (Szklarczyk et al., 2021). Using TRRUST (version 2) (Han et al., 2013). Epigenetic and genetic alterations are necessary for tumor 2018), we performed a transcriptional regulatory analysis to find the development (Takeshima and Ushijima, 2019). Alterations in the transcription factors regulating the identified genes. epigenome during tumorigenesis include global changes in histone modification marks (Fraga et al., 2005; Hosseini and Minucci, 2017), To assess whether the methylated probes identified in this study could deregulation of noncoding RNA networks (Liz and Esteller, 2016), be useful for diagnostic classification of hDLBCL, RFE was done to global DNA methylation loss and CpG promoter islands select the top 200 important features for classification. This was to hypermethylation of tumor suppressor genes (Baylin and Jones, accommodate for the highly reduced gene number in the human 2016). Few studies have been done to characterize the epigenetic dataset to be used for this assessment. The GSE28094 dataset was landscape of cDLBCL and these have been focused on DNA retrieved from the GEO database (Fernandez et al., 2012). This dataset methylation. These included DNA methylation studies that focused on is a human DNA methylation dataset comprising 424 normal tissue single gene (Sato et al., 2014; Fujiwara-Igarashi et al., 2014; Tomiyasu samples, 1054 tumor samples and 150 non-cancerous disorders et al., 2014; Ferraresso et al., 2014), and genome-wide (Ferraresso et samples. The beta values of 1505 CpG sites corresponding to 808 al., 2017; Aresu et al., 2019; Hsu et al., 2021). The full understanding of genes were profiled in the dataset. We extracted the profiles for 49 the molecular mechanisms underlying cDLBCL is still incomplete. This DLBCL and 49 healthy normal leukocytes from the dataset. The study aims to explore the utility of DNA methylation data for available genes out of the 200 top genes identified from RFE were diagnostic classification of DLBCL in dogs. The identified classification thereafter used as classification features. Clustering of these 98 human features would also be assessed for diagnostic classification of DLBCL samples was then performed using factoextra package (Kassambara in humans. and Mundt, 2020). Interaction analysis and transcriptional regulatory analysis of the genes were performed using STRING database (version 2. Experimental 11.5) (Szklarczyk et al., 2021) and TRRUST (version 2) (Han et al., 2018) respectively. Python and R scripts used in this study’s analyses Dataset are available in https://github.com/ZEB-LASU/cDLBCL The GSE94913 cDLBCL DNA methylation dataset was retrieved from the GEO database (Ferraresso et al., 2017). This dataset is composed of 3. Results 37 cDLBCL samples and 7 control lymph node samples. This dataset is The differential methylation analysis between the cDLBCL and control a microarray of 37479 CpG and coding sequences (CDS) probes, with lymph node groups identified 1701 differentially methylated probes. standard microarray data normalization and processing applied. 100% of the 44 samples were accurately clustered using these 1701 significantly methylated probes (fig. 1). The top 20 probes for Differential methylation analysis classifying samples selected by RFE are shown in Table 1. Annotation Differential methylation analysis was performed between the 37 9| This journal is © The Nigerian Young Academy 2022 A nn a ls of Science and Technology 2023 Vol. 8 (1): 8-15 Exploring DNA methylation data Giwa and Adu., 2023 on the Ensembl genome browser (CanFam 3.1) identified these to To assess the diagnostic classification utility of the methylated genes include eleven protein-coding genes, three non coding RNAs, and six in hDLBCL, 11 CpG sites of six genes (ERBB4, IGF2, PGF, PITX2, TJP1, unidentified genes (Table 1). These were used in ML training and test PDGFA) was used as features for clustering the 98 samples obtained set building. The training set was built on 30 samples randomly from the GSE28094 dataset. 95% of the samples were accurately selected by the train_test_split Scikit-learn function, comprising of 25 clustered into their respective groups (fig. 2). The 5 misclassified cDLBCL samples and 5 control lymph node samples. The testing set samples were however in a separate cluster group even though placed was made up of 14 samples selected by train_test_split Scikit-learn in the same branch with the normal leukocyte tissues (fig. 2). A function, comprising of 12 cDLBCL samples and 2 control lymph node supervised machine learning model could not be constructed due to samples. The 5-fold cross-validation accuracy obtained was 100%. the difference in platforms and the small number of genes profiled in Evaluation of the SVM model on the testing set resulted in an accuracy the available hDLBCL dataset. The interaction analysis between the six of 100% of correctly predicted samples (Table 2). For the interaction genes identified interactions between four of the six genes (PDGFA, and transcriptional regulatory analysis, no interaction and PGF, ERBB4, IGF2) (fig. 3). The transcriptional regulatory analysis transcriptional regulatory relationships were returned. revealed WT1 regulating the genes IGF2 and PDGFA. Fig. 1. Clustering of the cDLBCL and control lymph node samples using all 1701 significant methylation probes. The x-axis contains the samples and which are labelled for clear identification. Identified were two principal clusters with the 7 control lymph node samples all grouped in the left cluster (black rectangle box) while the 37 cDLBCL samples were all grouped in the right cluster (red rectangle box). DNA methylation data for diagnostic classification of DLBCL in 4. Discussion Applying machine learning classification techniques to cDLBCL DNA methylation data can give good results and thus can be beneficial The differential methylation analysis revealed expectedly that there Dysregulated expression of some of these identified top genes have were differences in methylation between cDLBCL and the control been associated with several cancers including RAB7A in gastric samples. Indeed, aberrant DNA methylation is characteristic of cancer cancer (Liu et al., 2020), RAB7A in breast cancer (Xie et al., 2019), (Dong et al., 2014), and cDLBCL (Ferraresso et al., 2017; Fujiwara- LEF1 in melanoma (Kim et al., 2020), IGF2 in hepatocellular Igarashi et al., 2014; Hsu et al., 2021; Sato et al., 2018). The accuracy of carcinoma (Ma et al., 2019) and adrenocortical carcinoma (Pereira et al., 2019), ST8SIA2 in lung cancer (Hao et al., 2019), DDIT4 in acute the clustering despite using all 1701 significant probes substantiated myeloid leukemia (Cheng et al., 2020) and in ovarian carcinoma the differential methylation analysis result and indicated their (Chang et al., 2018), and PPM1H in colon adenocarcinoma (Sugiura et possible utility for classification. The high accuracy obtained by the ML al., 2008) and in pancreatic cancer (Zhu et al., 2016). These genes model (Table 2) constructed using 20 significant probes (features) could be potential drivers of cDLBCL in Dogs. Aberrant methylation of these genes could affect their expression. They can also serve as demonstrate the validity of these CpGs for diagnostic identification of possible therapeutic targets upon further studies and functional DLBCL in dogs. As a limitation, the testing set was small, requiring studies of these identified genes should elucidate our understanding validation in larger sample cohorts. We were able to explore the use of DNA methylation data for diagnostic classification of DLBCL in Dogs. This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 10 Exploring DNA methylation data Giwa and Adu., 2023 Table 1. Twenty probes selected by recursive feature elimination and used as features for machine learning classification Probe ID adj-p-value Gene name GT_chr6:31498360-31499687_296737 2.40686e-05 TNP2* GT_CpG_102:chr20:2792244-2793806_10609 5.00947e-19 RAB7A* GT_CpG_133:chr8:34699507-34700781_125659 4.57457e-07 JKAMP* GT_CpG_199:chr32:28632693-28635175_287611 1.84638e-06 LEF1* GT_CpG_208:chr10:6070361-6072996_308205 2.76684e-22 PPM1H* GT_CpG_217:chr3:47653366-47655868_330479 4.13005e-17 ST8SIA2* GT_CpG_31:chr4:22936675-22937328_508162 6.84959e-12 DDIT4* GT_CpG_35:chr25:15429047-15429606_825639 7.52139e-17 SGCG* GT_CpG_42:chr18:46302185-46302823_622168 5.85738e-12 IGF2* GT_CpG_57:chr22:8708474-8709185_748846 3.30485e-08 DGKH* GT_CpG_62:chr6:17872282-17873207_782319 1.75655e-12 ENSCAFT00000077585** GT_CpG_72:chr10:55372722-55373751_851800 1.48866e-14 ENSCAFT00000065559** GT_CpG_82:chr19:23852100-23852984_905754 0.00054 ENSCAFT00000089251** GT_CpG_90:chr1:104706255-104707295_948211 3.10691e-19 ENSCAFT00000044531* GT_CpG_103:chr9:5445423-5446725_17310 8.35098e-18 GT_CpG_33:chr23:33251449-33252124_530445 2.76684e-22 GT_CpG_43:chr15:45025176-45025844_631179 6.43581e-19 GT_CpG_69:chr11:62398353-62399319_830114 1.94461e-17 GT_CpG_77:chr27:31835085-31836199_881657 1.55622e-14 GT_CpG_99:chr6:38018350-38019848_992235 1.37342e-07 Shown are the probe identifiers, their adjusted p-values, and corresponding gene names. *protein coding genes. **long non-coding RNA gene. Blanks are genes without annotation names Table 2: Machine learning classification results of the testing set using SVM model Class Precision Recall cDLBCL 1.0 1.0 Control lymph node 1.0 1.0 Accuracy 100% (14/14) 11| This journal is © The Nigerian Young Academy 2022 A nn als of Science and Technology 2023 Vol. 8 (1): 8-15 Exploring DNA methylation data Giwa and Adu., 2023 Fig. 2. Clustering of the 98 DLBCL and leukocyte samples of the GSE28094 human dataset. The x-axis contains the samples, with the DLBCL samples labelled with D and leukocyte samples labelled with HL. Identified were two main clusters with the DLBCL samples predominantly clustered in the right cluster (red rectangle box) while the leukocyte samples were predominantly clustered in the left cluster (black rectangle box). 5 DLBCL samples were misclassified. Fig. 3. Gene interaction analysis of the six (ERBB4, IGF2, PGF, PITX2, TJP1, PDGFA) genes used for clustering the hDLBCL samples. understanding of DLBCL disease mechanisms in dogs. The be potentially used for diagnostic identification of DLBCL in humans unavailability of external datasets to further validate this CpG upon further validation. WT1 was reported to have transcriptionally signature indicates the need for more investigative studies to regulate IGF2 and PDGFA, and has been associated with several characterize the epigenetic landscape of DLBCL in dogs. The cancers (Qi et al., 2015; Zhang et al., 2020). Diagnostic classification of unsupervised model based on 11 CpGs of six genes (ERBB4, IGF2, PGF, tumors using DNA methylation data has been demonstrated in PITX2, TJP1, PDGFA) was able to discriminate hDLBCL samples and multiple human cancers (Capper et al., 2018; Liu et al., 2020; Giwa et leukocyte samples with high accuracy (fig 2). This reinforces the al., 2021). It could therefore be possible to develop diagnostic molecular and clinical similarity between cDLBCL and hDLBCL biomarkers that are clinically useful in canine and human DLBCL. (Richards et al., 2013; Aresu et al., 2019). These six genes could thus This journal is © The Nigerian Young Academy 2023 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 12 Giwa and Adu., 2023 Exploring DNA methylation data 5. Conclusion Chen, Y., Hesla, A.C., Lin, Y., Ghaderi, M., Liu, M., Yang, C., Zhang, Y., This exploratory study demonstrates the utility of DNA methylation Tsagkozis, P., Larsson, O., and Haglund, F., (2020), Transcriptome data for diagnostic classification of DLBCL in dogs. DNA methylation profiling of Ewing sarcomas - treatment resistance pathways and IGF- data-based diagnostics for tumor classification is becoming dependency. Molecular Oncology, 14(5):1101-1117. mainstream. Therefore, characterizing the methylation landscape of DLBCL in dogs should have clinical benefit. It will also further Cheng, Z., Dai, Y., Pang, Y., Jiao, Y., Liu, Y., Cui, L., Quan, L., Qian, T., Zeng, understanding of the disease mechanisms and aid therapeutic T., Si, C., Huang, W., Chen, J., Pang, Y., Ye, X., Shi, J., and Fu, L., (2020), development as survival rate of DLBCL in dogs is still low. Further Up-regulation of DDIT4 predicts poor prognosis in acute myeloid exploration of this data type for potential biomarker discovery in leukaemia. Journal of Cellular and Molecular Medicine, 24(1):1067- cDLBCL is necessary. Declaration of Conflict of Interests Dong, Y., Zhao, H., Li, H., Li, X., and Yang, S., (2014), DNA methylation as No conflict of interests to declare an early diagnostic marker of cancer (Review). Biomedical Reports, 2(3):326-330. Authors’ Contributions Fernandez, A.F., Assenov, Y., Martin-Subero, J.I., Balint, B., Siebert, R., Conception: [AG] Taniguchi, H., Yamamoto, H., Hidalgo, M., Tan, A.C., Galm, O., Ferrer, I., Design: [AG] Sanchez-Cespedes, M., Villanueva, A., Carmona, J., Sanchez-Mut, J.V., Execution: [OA, AG] Berdasco, M., Moreno, V., Capella, G., Monk, D., Ballestar, E., Ropero, S., Interpretation: [OA, AG] Martinez, R., Sanchez-Carbayo, M., Prosper, F., Agirre, X., Fraga, M.F., Writing the paper: [OA, AG] Graña, O., Perez-Jurado, L., Mora, J., Puig, S., Prat, J., Badimon, L., Puca, A.A., Meltzer, S.J., Lengauer, T., Bridgewater, J., Bock, C., and Esteller, M., (2012), A DNA methylation fingerprint of 1628 human samples. References Genome Research, 22(2):407-419. Aresu, L., Martini, V., Rossi, F., Vignoli, M., Sampaolo, M., Aricò, A., Ferraresso, S., Bresolin, S., Aricò, A., Comazzi, S., Gelain, M.E., Riondato, Laganga, P., Pierini, A., Frayssinet, P., Mantovani, R., and Marconato, L., F., Bargelloni, L., Marconato, L., Kronnie, Gt., and Aresu, L., (2014), (2015), Canine indolent and aggressive lymphoma: clinical spectrum Epigenetic silencing of TFPI-2 in canine diffuse large B-cell lymphoma. with histologic correlation. Veterinary and Comparative Oncology, PLoS One, 9(4):e92707. 13(4):348-362. Ferraresso, S., Aricò, A., Sanavia, T., Da Ros, S., Milan, M., Cascione, L., Aresu, L., (2016), Canine Lymphoma, More than a morphological Comazzi, S., Martini, V., Giantin, M., Di Camillo, B., Mazzariol, S., diagnosis: what we have learned about diffuse large B-cell lymphoma. Giannuzzi, D., Marconato, L., and Aresu, L., (2017), DNA methylation Frontiers in Veterinary Science, 3:77. profiling reveals common signatures of tumorigenesis and defines epigenetic prognostic subtypes of canine Diffuse Large B-cell Aresu, L., Ferraresso, S., Marconato, L., Cascione, L., Napoli, S., Gaudio, Lymphoma. Scientific Reports, 7(1):11591. E., Kwee, I., Tarantelli, C., Testa, A., Maniaci, C., Ciulli, A., Hillmann, P., Bohnacker, T., Wymann, M.P., Comazzi, S., Milan, M., Riondato, F., Fraga, M.F., Ballestar, E., Villar-Garea, A., Boix-Chornet, M., Espada, J., Rovere, G.D., Giantin, M., Giannuzzi D., and Bertoni, F., (2019), New Schotta, G., Bonaldi, T., Haydon, C., Ropero, S., Petrie, K., Iyer, N.G., molecular and therapeutic insights into canine diffuse large B-cell Pérez-Rosado, A., Calvo, E., Lopez, J.A., Cano, A., Calasanz, M.J., Colomer, lymphoma elucidates the role of the dog as a model for human disease. D., Piris, M.A., Ahn, N., Imhof, A., Caldas, C., Jenuwein, T., and Esteller, Haematologica, 104(6):e256-e259. M., (2005), Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human cancer. Nature Genetics, Baylin, S.B., and Jones, P.A., (2016), Epigenetic determinants of cancer. 37(4):391-400. Cold Spring Harbor Perspectives in Biology, 8(9):a019505. Fujiwara-Igarashi, A., Goto-Koshino, Y., Mochizuki, H., Sato, M., Fujino, Capper, D., Jones, D.T.W., Sill, M., Hovestadt, V., Schrimpf, D., et al. Y., Ohno, K., and Tsujimoto, H., (2014), Inhibition of p16 tumor (2018), DNA methylation-based classification of central nervous suppressor gene expression via promoter hypermethylation in canine system tumours. Nature, 555(7697):469-474. lymphoid tumor cells. Research in Veterinary Science, 97(1):60-63. Chang, B., Meng, J., Zhu, H., Du, X., Sun, L., Wang, L., Li, S., and Yang, G., Giannuzzi, D., Marconato, L., Cascione, L., Comazzi, S., Elgendy, R., (2018), Overexpression of the recently identified oncogene REDD1 Pegolo, S., Cecchinato, A., Bertoni, F., Aresu, L., and Ferraresso, S., correlates with tumor progression and is an independent unfavorable (2019), Mutational landscape of canine B-cell lymphoma profiled at prognostic factor for ovarian carcinoma. Diagnostic Pathology, single nucleotide resolution by RNA-seq. PLoS ONE, 14(4): e0215154. 13(1):87. Chang, C.C., and Lin, C.J., (2011), LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:27 13| This journal is © The Nigerian Young Academy 2022 Annals of Science and Technology 2023 Vol. 8 (1): 8-15 Giwa and Adu., 2023 Exploring DNA methylation data Giwa, A., Rossouw, S.C., Fatai, A., Gamieldien, J., Christoffels, A., and Martini, V., Aresu, L., Riondato, F., Marconato, L., Cozzi, M., Stefanello, Bendou, H., (2021), Predicting amplification of MYCN using CpG D., and Comazzi, S., (2019), Prognostic role of non-neoplastic methylation biomarkers in neuroblastoma. Future Oncology, lymphocytes in lymph node aspirates from dogs with diffuse large B- 17(34):4769-4783. cell lymphoma treated with chemo-immunotherapy. Research in Veterinary Science, 125:130-135. Han, H., Cho, J.W., Lee, S., Yun, A., Kim, H., Bae, D., Yang, S., Kim, C.Y., Lee, M., Kim, E., Lee, S., Kang, B., Jeong, D., Kim, Y., Jeon, H.N., Jung, H., Mizuno, T., Kato, Y., Kaneko, M.K., Sakai, Y., Shiga, T., Kato, M., Tsukui, Nam, S., Chung, M., Kim, J.H., and Lee, I., (2018), TRRUST v2: an T., Takemoto, H., Tokimasa, A., Baba, K., Nemoto, Y., Sakai, O., and expanded reference database of human and mouse transcriptional Igase, M., (2020), Generation of a canine anti-canine CD20 antibody for regulatory interactions. Nucleic Acids Research, 46(D1):D380-D386. canine lymphoma treatment. Scientific Reports, 10(1):11476. Hao, J., Zeltz, C., Pintilie, M., Li, Q., Sakashita, S., Wang, T., Cabanero, M., Mudaliar, M.A., Haggart, R.D., Miele, G., Sellar, G., Tan, K.A., Goodlad, Martins-Filho, S.N., Wang, D.Y., Pasko, E., Venkat, K., Joseph, J., J.R., Milne, E., Vail, D.M., Kurzman, I., Crowther, D., and Argyle, D.J., Raghavan, V., Zhu, C.Q., Wang, Y.H., Moghal, N., Tsao, M.S., and Navab, (2013), Comparative gene expression profiling identifies common R., (2019), Characterization of distinct populations of carcinoma- molecular signatures of NF-κB activation in canine and human diffuse associated fibroblasts from non-small cell lung carcinoma reveals a large B cell lymphoma (DLBCL). PLoS One, 8(9):e72591. role for ST8SIA2 in cancer cell invasion. Neoplasia, 21(5):482-493. Pedregosa F, Varoquaux G, Gramfort A, Michel V., Thirion, B., Grisel, O., Hosseini, A., and Minucci, S., (2017), A comprehensive review of Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., lysine-specific demethylase 1 and its roles in cancer. Epigenomics, Passos, A., Cournapeau, D., Brucher, M., Perrot, M., and Duchesnay, E., 9(8):1123-1142. (2011), Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12:2825–2830. Hsu, C.H., Tomiyasu, H., Lee, J.J., Tung, C.W., Liao, C.H., Chuang, C.H., Huang, L.Y., Liao, K.W., Chou, C.H., Liao, A.T.C., and Lin, C.S., (2021), Pereira, S.S., Monteiro, M.P., Costa, M.M., Moreira, Â., Alves, M.G., Genome-wide DNA methylation analysis using MethylCap-seq in Oliveira, P.F., Jarak. I., and Pignatelli, D., (2019), IGF2 role in canine high-grade B-cell lymphoma. Journal of Leukocyte Biology, adrenocortical carcinoma biology. Endocrine, 66(2):326-337. 109(6):1089-1103. Qi, X.W., Zhang, F., Wu, H., Liu, J.L., Zong, B.G., Xu, C., Jiang, J., (2015), Kassambara, A., and Mundt, F., (2020), Factoextra R package. Wilms' tumor 1 (WT1) expression and prognosis in solid cancer https://CRAN.R-project.org/package=factoextra patients: a systematic review and meta-analysis. Scientific Reports, 5:8924. Kim, G.H., Fang, X.Q., Lim, W.J., Park, J., Kang, T.B., Kim, J.H., and Lim, J.H., (2020), Cinobufagin suppresses melanoma cell growth by Richards, K.L., Motsinger-Reif, A.A., Chen, H.W., Fedoriw, Y., Fan, C., inhibiting LEF1. International Journal of Molecular Sciences, Nielsen, D.M., Small, G.W., Thomas, R., Smith, C., Dave, S.S., Perou, C.M., 21(18):6706. Breen, M., Borst, L.B., and Suter, S.E., (2013), Gene profiling of canine B-cell lymphoma reveals germinal center and postgerminal center Klimiuk, P., Łopuszynski, W., Bulak, K., and Brzana, A., (2021), subtypes with different survival times, modeling human DLBCL. Evaluation of the proliferative activity of diffuse large b-cell lymphoma Cancer Research, 73(16):5029-5039. (DLBCL) in dogs with respect to patient eligibility for anthracycline- based chemotherapy. Animals (Basel), 11(4):1183. Riondato, F., and Comazzi, S., (2021), Flow Cytometry in the Diagnosis of Canine B-Cell Lymphoma. Frontiers in Veterinary Science, Liu, H., Xu, J., Yao, Q., Zhang, Z., Guo, Q., and Lin, J., (2020), Rab7 is 8:600986. associated with poor prognosis of gastric cancer and promotes proliferation, invasion, and migration of gastric cancer cells. Medical Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, Science Monitor, 26:e922217. G.K., (2015), limma powers differential expression analyses for RNA- sequencing and microarray studies. Nucleic Acids Research, 43(7):e47. Liu, M.C., Oxnard, G.R., Klein, E.A., Swanton, C., Seiden, M.V.; CCGA Consortium, (2020), Sensitive and specific multi-cancer detection and Sato, M., Mochizuki, H., Goto-Koshino, Y., Fujiwara-Igarashi, A., localization using methylation signatures in cell-free DNA. Annals of Takahashi, M., Fujino, Y., Ohno, K., and Tsujimoto, H., (2014), Oncology, 31(6):745-759. Hypermethylation of the death-associated protein kinase CpG island in canine B-cell lymphoid tumors. Veterinary Immunology and Liz, J., and Esteller, M., (2016), lncRNAs and microRNAs with a role in Immunopathology, 161(3-4):222-231. cancer development. Biochimica et Biophysica Acta, 1859(1):169–176. Sato, M., Mochizuki, H., Goto-Koshino, Y., Fujiwara-Igarashi, A., Ma, Y., Chen, Y., Chen, L., Liu, Z., Ieong, M.L., Gao, F., and Huang, W., Takahashi, M., Ohno, K., and Tsujimoto, H., (2018), Prognostic (2018), Promotion of Insulin-like growth factor II in cell proliferation significance of hypermethylation of death-associated protein kinase and epithelial-mesenchymal transition in hepatocellular carcinoma. (DAPK) gene CpG island in dogs with high-grade B-cell lymphoma. Journal of Cancer Research and Therapeutics, 14(4):844-850. Veterinary and Comparative Oncology, 16(3):409-415. Annals of Science and Technology 2023 Vol. 8 (1): 8-15 | 14 This journal is © The Nigerian Young Academy 2023 Exploring DNA methylation data Giwa and Adu., 2023 Sirivisoot, S., Kasantikul, T., Techangamsuwan, S., Radtanakatikanon, Tomiyasu, H., Fujiwara-Igarashi, A., Goto-Koshino, Y., Fujino, Y., Ohno, A., Chen, K., Lin, T.Y., and Rungsipipat, A., (2022), Evaluation of 41 K., and Tsujimoto, H., (2014), Evaluation of DNA methylation profiles single nucleotide polymorphisms in canine diffuse large B-cell of the CpG island of the ABCB1 gene in dogs with lymphoma. American lymphomas using MassARRAY. Scientific Reports, 12(1):5120. Journal of Veterinary Research, 75(9):835-841. Sugiura, T., Noguchi, Y., Sakurai, K., and Hattori, C., (2008), Protein Xie, J., Yan, Y., Liu, F., Kang, H., Xu, F., Xiao, W., Wang H., and Wang, Y., phosphatase 1H, overexpressed in colon adenocarcinoma, is (2019), Knockdown of Rab7a suppresses the proliferation, migration, associated with CSE1L. Cancer Biology and Therapy, 7(2):285-292. and xenograft tumor growth of breast cancer cells. Bioscience Reports, 39(2):BSR20180480. Szklarczyk, D., Gable, A.L., Nastou, K.C., Lyon, D., Kirsch, R., Pyysalo, S., Doncheva, N.T., Legeay, M., Fang, T., Bork, P., Jensen, L.J., and von Zandvliet, M., (2016), Canine lymphoma: a review. Veterinary Mering, C., (2021), The STRING database in 2021: customizable Quarterly, 36(2):76-104. protein-protein networks, and functional characterization of user- uploaded gene/measurement sets. Nucleic Acids Research, Zhang, Y., Yan, W.T., Yang, Z.Y., Li, Y.L., Tan, X.N., Jiang, J., Zhang, Y., and 49(D1):D605-D612. Qi, X.W., (2020), The role of WT1 in breast cancer: clinical implications, biological effects and molecular mechanism. Takeshima, H., and Ushijima, T., (2019), Accumulation of genetic and International Journal of Biological Sciences, 16(8):1474-1480. epigenetic alterations in normal cells and cancer risk. NPJ Precision Oncology, 3:7. Zhu, H., Qin, H., Li, D.M., Liu, J., and Zhao, Q., (2016), Effect of PPM1H on malignant phenotype of human pancreatic cancer cells. Oncology Reports, 36(5):2926-2934. 15| This journal is © The Nigerian Young Academy 2022 Annals of Science and Technology 2023 Vol. 8 (1): 8-15

Journal

Annals of Science and Technologyde Gruyter

Published: Jun 1, 2023

Keywords: Biomarkers; Classification; DNA methylation; Dogs; Machine learning

There are no references for this article.