Supplementary Figure 1. Exploratory data analysis of NSCs and GSCs samples using PCA based on all the : (a) PCA plot for all samples, (b) PCA plot of all samples after removal of outlier sample (SRR9200898_PE).

Supplementary Figure 2. Dendrogram from H-Clustering analysis depicting the clusters of NSC and GSC samples based on all genes.

Supplementary Table S1: List of top 25 upregulated genes with their roles in normal cells and Glioma.

ENTREZ ID Full name Function in normal Function in glioma Referenc pathology es

S100B S100 This encodes a Shorter survival (Frauchig Calcium with functions in period is associated er et al., Binding neurite extension, with high levels of 2019; Protein B proliferation of melanoma serum S100B in Gao et cells, stimulation of Ca2+ glioma patients. It al., 2018; fluxes, inhibition of PKC- promotes glioma Holla et mediated phosphorylation, growth by Tumor- al., 2016; astrocytosis and axonal associated Sorci et proliferation, and macrophages al., 2013) inhibition of microtubule (TAMs) assembly. chemoattraction through upregulation of CCL2 and thus S100B inhibitors are potential drugs for glioma therapy.

PLP1 Proteolipid This gene encodes a PLP1 association (Luo et Protein 1 transmembrane with glioma al., 2019; proteolipid protein present invasion. Needs Miyamot in which play a further investigation o et al., role in maintaining the to determine a 2012; compaction and significant Ocklenbu stabilization of myelin functional role. rg et al., sheaths, as well as in 2019) development and survival of oligodendrocytes.

S100A6 S100 Involved in the regulation S100A6 are (Donato Calcium of a number of cellular expressed in a small et al., Binding processes such as cell subset of cancer 2017, p. Protein A6 cycle progression and stem cells and that 6; Harris differentiation. its overexpression is et al., positively correlated 2008) to tumor grade.

PMP2 Peripheral The protein encoded by PMP2 is associated (Hong et Myelin this gene is a component of with the stemness of al., 2016, Protein 2 the myelin sheaths of the GBM and a p. 2; Vital peripheral nervous system. discriminatory et al., Mutation in this gene is biomarker between 2010) shown to be the cause of gliomas and dominant demyelinating gliosarcomas. CMT neuropathy. However, individual gene studies needed.

ITM2A Integral In vivo studies have Little known. In (Deleersn Membrane suggested the protein’s other cancers it ijder et Protein 2A role in osteo- and prevents al., 1996; chondrogenic proliferation and Zhang et differentiation. invasion. More al., 2021, studies are required. p. 2) GPNMB Glycoprotein GPNMB may be Overexpression of (Liguori nonmetastati responsible for prolonging GPNMB is et al., c melanoma cell survival and tumor associated with a 2020; protein B growth while increasing poor prognosis in Ono et metastatic potential of glioblastoma al., 2016) malignant tumors. patients. GPNMB promotes glioma growth via Na+/K+- ATPase α subunits which can be exploited as a novel therapeutic target for the treatment of GBM.

THY1/ Thy-1 Cell The encoded protein is CD90 is an (Avril et CD90 Surface involved in cell adhesion established stem cell al., 2017, Antigen and cell communication in marker and p. 90; cells of the immune and expressed in GBM- Sauzay et nervous systems. This associated stromal al., 2019) gene may function as a cells (GASCs) and tumor suppressor in mesenchymal stem nasopharyngeal cell-like pericytes, carcinoma. indicative of the heterogenous nature of GBM. It is therefore a good therapeutic target. AZGP1 Alpha-2- Stimulates lipid It is one of the (Huang et Glycoprotein degradation in adipocytes overexpressed al., 2013; 1, Zinc- and causes the extensive genes in the GBM Osti et Binding fat losses associated with Extracellular al., 2019) some advanced cancers. Vesicles protein signature. More studies are needed.

RPS4Y1 Ribosomal It is a Y gene Associated with (Khosrav Protein S4 Y- having an important role in GBM but requires i et al., Linked 1 prostate cancer. more studies 2014; Wipfler et al., 2018)

MATN2 Matrilin 2 This gene encodes a Associated with (Varga et extracellular matrix GBM, however, al., 2010; (ECM) protein that more studies are Zhang et balances the needed. al., 2014) communication between ECM and epithelial cells. GAS7 Growth This protein is expressed in GAS7 has been (Ju et al., Arrest terminally differentiated detected at low to 1998, p. Specific 7 brain cells such as in moderate levels and 7; mature cerebellar Purkinje is not deemed Tatenhor neurons. It plays an prognostic for st et al., important role in neuronal glioma. More 2005; development. studies are required. You and Lin- Chao, 2010, p. 7)

KRBOX1 KRAB Box KRBOX1 are DNA- Has not been (Park et Domain binding repressors of associated yet, thus, al., 2017) transcription. The targets Containing 1 requires further of this protein remain largely unknown. studies.

OLIG1 Oligodendroc OLIG1 (Oligodendrocyte Lower expression of (Bouvier yte Transcription Factor 1) is a Olig1 observed in et al., Transcription protein-coding gene glioblastoma. Olig1 2003, p. Factor 1 associated with transcription factor 1; Dai et Oligodendroglioma and is required for al., 2015; Grade Iii Astrocytoma. It maturation of Wu et al., is involved in pathways oligodendrocyte 2012) such as Neural Crest progenitors. Differentiation and Neural Stem Cells and Lineage- specific Markers. PRR34-AS1 PRR34 PRR34-AS1 (PRR34 No associated found (Kesherw Antisense Antisense RNA 1) is an yet, thus, need more ani et al., RNA 1 RNA Gene, and is studies. 2020) affiliated with the lncRNA class. It has been identified to play a role in medulloblastoma.

SOX10 SRY-Box It encodes a protein that Sox10 has been (Ferletta Transcription acts as a identified as a et al., Factor 10 nucleocytoplasmic shuttle master-regulator of 2007, p. protein and is important for the RTKI subtype in 1; neural crest and peripheral GBM. Low-grade Pozniak nervous system gliomas have a et al., development. Among its higher expression of 2010; related pathways are Sox10 compared to Rehberg Neural Crest high-grade glioma. et al., Differentiation and Neural 2002; Wu Stem Cells and Lineage- et al., specific Markers. 2020)

ENSG0000015 N/A N/A N/A N/A 4553.12 (unknown)

TSPAN7 7 This glycoprotein and may TSPAN7 has been (Bassani have a role in the control of identified as a good et al., neurite outgrowth. Among prognostic 2012; its related pathways are biomarker, Wuttig et dysregulation of associated with a al., 2012) transcription in cancer and longer disease-free chemical transmission survival and tumor- across synapses. specific survival.

CDH9 Cadherin 9 This gene encodes a type II CDH9 has a role in (Turaga classical cadherin from the the transition of and cadherin superfamily, CSCs to Endothelial Lathia, integral membrane Cells (ECs). More 2016; that mediate studies are required. Wang et calcium-dependent cell- al., 2019, cell adhesion. It has been p. 9) associated to play a role in the development of autism spectrum disease in cerebellar pathways.

MT1F Metallothion Metallothioneins have a MT1F has shown to (Lu et al., ein 1F high content of cysteine play a role in 2003; residues that bind various imparting Mehrian- heavy metals; these chemoresistance in Shai et proteins are glioblastoma cells. al., 2015; transcriptionally regulated It may also be a Yan et by both heavy metals and significant al., 2012) glucocorticoids. It may be prognostic an essential tumor-growth biomarker to suppressor in differentiate short- hepatocellular carcinoma and long-term and colon cancer. patient survival. NTRK2 Neurotrophic This gene encodes a NTRK2 is known to (Jones et Receptor member of the rearrange or fuse al., 2019; Tyrosine neurotrophic tyrosine with other genes Pattwell Kinase 2 receptor kinase (NTRK) such as BCR and is et al., family that plays a role in a associated with 2020a, signalling pathway leading driving 2020b; to cell survival and tumorigenesis and Torre et differentiation. increasing al., 2020) aggressiveness of glioblastoma.

FCRLA Fc Receptor This protein may also be No known studies (Capone Like A involved in the performed yet and et al., development of thus, further 2016; lymphomas. It has several investigation Santiago known spliced transcript needed. et al., isoforms. 2011)

SLCO4A1 Solute SLCO4A1 is involved in Has still not been (Ban et Carrier several pathways such as associated with al., 2017; Organic transport of vitamins, GBM and requires Sun et al., Anion nucleosides, and related further studies. One 2018) Transporter molecules and transport of study found it to be Family glucose and other sugars, expressed in Member 4A1 bile salts and organic differentially acids, metal ions and methylated region amine compounds. Its role GBM cells. has been established in several cancers including colorectal cancer. PDLIM3 PDZ And The protein encoded by Associated but (Kim et LIM Domain this gene contains a PDZ needs further al., 2013; 3 domain and a LIM domain, studies. Maurin et indicating that it may be al., 2009) involved in cytoskeletal assembly. It is known to be involved in cardiac and muscular disease pathogenesis.

MIA MIA SH3 Diseases associated with MIA mRNA (Hau et Domain MIA include Melanoma. expression has been al., 2002; Containing Among its related observed in Poser et pathways are Neural Crest melanoma and al., 2004) Differentiation. It is glioma tumor cells responsible for growth but not in non- inhibition in melanoma, glioma CNS tumors neuroectodermal tumors, making it a potential including gliomas. diagnostic biomarker. More studies are needed.

LY96 Lymphocyte This gene encodes a Associated with (Dou et Antigen 96 protein which associates poor survival in al., 2013; with toll-like receptor 4 on GBM, however, Moreno the cell surface and confers needs further et al., responsiveness to studies. 2021; lipopolysaccyaride (LPS), Rajarama thus providing a link n et al., between the receptor and 2009) LPS signaling.

Supplementary Table S2: List of top 25 upregulated genes with their roles in normal cells and gliomas.

ENTREZ Full name Function Function in glioma Referenc ID es

S100A11 S100 Calcium This protein plays a role in S100A11 is a prognostic (Mori et Binding the functions of motility, marker for GBM with al., 2004, Protein A11 invasion, tubulin high expression p. 1; Tu polymerization. Mutations indicating poor outcome. et al., and chromosomal It’s overexpression 2019, p. rearrangements of this gene promotes cell 11) may play a role in tumor proliferation, epithelial- metastasis. mesenchymal transition (EMT), migration and invasion of glioma stem cells (GSCs), whereas its knockdown has shown to inhibit these roles.

UBB Ubiquitin B This gene encodes Associated with the (Kedves ubiquitin, that is involved regulation of several et al., in the maintenance of pathways of GBM, 2017; chromatin structure, however their specific Scholz et regulation of gene function remains al., 2020) expression, and stress unknown. More studies response. Its repression has are required. been linked with ovarian, uterine and endometrial cancers.

XLOC_02 Novel gene N/A N/A N/A 9252 MGST1 Microsomal This protein protects the Downregulation of (Arment Glutathione membranes of the MGST1 weakens cell o et al., S-Transferase endoplasmic reticulum and adhesion (may lead to 2017; 1 mitochondria from metastasis). However, no Bräutiga oxidative stress. This gene specific association with m et al., is also essential for GBM has been found, 2018) embryonic development and thus, more studies and hematopoiesis in are required. vertebrates.

SERPINF Serpin Family It is a strong inhibitor of Loss of expression of this (Guan et 1 F Member 1 angiogenesis. It is a gene is involved in al., 2004; neurotrophic factor glioma proliferation. Xu et al., involved in neuronal 2017) differentiation in retinoblastoma cells.

SPARCL SPARC Like It is a tumor suppressor This gene is a potential (Gagliard 1 1 gene in several types of therapeutic biomarker for i et al., tumor while it may also be GBM as it may be 2020, an oncogene in other types. involved in driving the 2017) growth and infiltration of GSCs while also promoting angiogenesis.

MFAP5 Microfibril MFAP5 encode One of the genes of a (Cheng et associated extracellular matrix signature gene set al., 2012; protein 5 proteins that play important involved in epithelial- Wu et al., roles in bone, blood mesenchymal transition 2019, p. vessels, hemostasis and the in GBM. 5) immune system. Maybe upregulated in invasive types of cancers.

CRIP1 Cysteine Rich Seems to have a role in zinc Novel marker found to be (Ochock Protein 1 absorption and may expressed in glioma- a et al., function as an intracellular associated brain 2019; zinc transport protein. macrophages. Wang et Gene overexpressed in al., 2007) prostate and pancreatic cancers.

LUM Lumican Lumican regulates collagen Shown to be upregulated (Chakrav fibril organization, corneal in GSCs and may be arti, transparency, epithelial cell responsible for 2002; migration and tissue repair. chemoresistance of the Farace et tumor. al., 2015)

MYL9 Myosin Light The encoded protein binds It is a novel biomarker (Kruthik Chain 9 calcium and is activated by associated with poor a et al., myosin light chain kinase prognosis in GBM and a 2019, p. and regulates muscle good determinant of its 9; Luo et contractions. It’s role has aggressiveness as it’s al., 2014, been identified in highly expressed in p. 9; Tan colorectal and non-small recurrent GBM. and cell lung carcinomas. Chen, 2014, p. 9) RAB34 RAB34, This gene encodes a protein It’s overexpression is an (L. Sun et Member RAS belonging to the RAB indicator of poor al., 2018, Oncogene family of proteins, which prognosis and a lowered p. 34; Family are small GTPases overall survival of high- Wang et involved in protein grade GBM patients. al., 2015, transport. It is involved in p. 34; Xu hedgehog signaling and et al., also plays a role in 2018, p. adhesion, invasion and 34) migration of breast tumors.

TPM2 Tropomyosin This gene encodes beta- The downregulation of (Dube et 2 tropomyosin, and mainly this gene is responsible al., 2016; expressed in slow, type 1 for the infiltration of the Mitchell muscle fibers. This gene soft brain environment et al., may also be responsible for with the GBM tumor. 2019) the transformation of breast epithelial cells.

FEZF1 FEZ Family The encoded protein is FEZF1-AS1 is a (Shimizu Zinc Finger 1 thought to play a role in the potential prognostic and Hibi, embryonic migration of biomarker for GBM 2009; M. gonadotropin-releasing indicating poor prognosis Yu et al., hormone secreting- and of patients. It is a driver 2017; Yu monoaminergic- neurons of GBM proliferation et al., into the basal forebrain. and infiltration, and thus, 2018) is a key oncogene. It also overcomes apoptosis by inhibiting the PI3K/AKT pathways in GSCs. FBN2 Fibrillin 2 The protein encoded by this No known function in (Charbon gene is a component of gliomagenesis, thus, neau et connective tissue further studies are al., 2003; microfibrils and may be needed. van Loon involved in elastic fiber et al., assembly. It is exclusively 2020) expressed in the peripheral nerves. It may a play role in tumorigenesis via regulation of the TGF-β pathway.

SFTA1P surfactant SFTA1P is regarded as a No known role in GBM, (Huang associated 1, tumor suppressor in non- thus, studies are required. et al., lncRNA small cell lung cancer. May 2017; be potential targets for Zhang et therapy. al., 2017)

PDLIM4 PDZ And PDLIM4, a LIM domain PDLIM4 expression is (Feng et LIM Domain gene also known as RIL, is associated with al., 2010; 4 suspected to have tumor imparting GBM Ming et suppressor functions in radioresistance, and is al., 2017; myeloid diseases. correlated with the Tayrac et Hypermethylation of this aggressiveness and al., 2011; gene is a potential prognosis of brain Vanaja et biomarker in prostate and cancer. al., 2006) breast cancers. TUSC3 Tumor It is involved in cellular Downregulation of this (X. Yu et Suppressor magnesium uptake, protein gene is correlated with al., 2017, Candidate 3 glycosylation and higher grade of glioma, p. 3; embryonic development. as it is involved in Yuan et This gene is a candidate enhancing the al., 2018) novel tumor suppressor proliferation and gene. invasiveness of the tumor.

LOXL1 Lysyl It may be involved in Overexpression is (Kim et Oxidase Like developmental regulation, indicative of higher glade al., 2014; 1 senescence, tumor glioma. It is a great M. Li et suppression, cell growth biomarker that can be al., 2019; control, and chemotaxis. exploited in liquid biopsy Yu et al., to monitor tumor 2020) progression. It is also responsible for preventing apoptosis in gliomas.

SYT1 Synaptotagmi The synaptotagmins are It is a marker of the (Bacaj et n 1 Ca2+ sensors responsible Neural subtype of GBM. al., 2013; for vesicular trafficking More studies are required Zhang et and exocytosis, and to identify its role in al., 2020) triggering neurotransmitter gliomagenesis. release at the synapse. CAV2 Caveolin 2 The protein is involved in CAV2 is regulated by (Liu et essential cellular functions, miR-144 and enhances al., 2020, including signal glioma migration and p. 2; transduction, lipid invasion via EMT. Sowa, metabolism, cellular 2011) growth control and apoptosis. This protein may function as a tumor suppressor.

ARHGAP Rho GTPase Rap1 is a small GTPase Knockdown of this gene (Bruning 29 Activating that, through effectors, is responsible for the - Protein 29 regulates Rho GTPase decrease in migration of Richards signaling. It is a known GBM. on et al., promoter of metastasis in 2018; several types of cancers Kolb et such as breast and prostate. al., 2020)

OMD Osteomodulin It is a marker in the human No known association (Lin et dental pulp stem cells. It with gliomagenesis and al., 2019; may be involved in the thus, needs further Ninomiy regulation of osteogenesis. investigation. a et al., 2007)

XLOC_01 Novel gene Not known. Not known. 2003

SLC2A12 Solute Carrier SLC2A12 is involved in Not known thus further (Toyoda Family 2 regulating blood urate studies are needed. et al., Member 12 levels. It may be a good 2020; biomarker for gastric Zheng et cancer as its levels are al., 2020, lowered post-operation. p. 2)

XLOC_04 p53-regulated Not known. Not known. (Jain et 7367 lncRNAs al., 2016)