Full Text
<article class="scholarly-article">
<h2>Introduction</h2>
<p>Cancer metastasis remains the primary cause of mortality in oncological patients, representing a systemic stage of disease where localized treatments often fail. The transition of tumor cells from a stationary to an invasive phenotype involves a radical reorganization of cellular architecture and signaling priorities. While genomic alterations provide the template for oncogenesis, the functional execution of the metastatic program is largely mediated by post-translational modifications (PTMs), with protein phosphorylation being the most pervasive and dynamic [1, 2]. The advent of high-resolution mass spectrometry (MS) has revolutionized our ability to quantify these modifications at a global scale, allowing for the systematic interrogation of kinase-substrate networks that drive disease progression [4, 11].</p><p>Metastatic signaling is characterized by the "rewiring" of intracellular circuits, where conventional pathways are co-opted or bypassed to favor survival, motility, and colonization of distant organs [11, 25]. Recent studies have highlighted how specific kinases, such as Wee1 in glioblastoma [2] or DCLK1 in ovarian cancer [14], act as central hubs in these rewired networks. However, identifying the specific substrates and the downstream effector pathways remains a significant challenge due to the transient nature of kinase-substrate interactions and the complexity of the phosphoproteome [4, 6]. Quantitative phosphoproteomics, particularly when combined with sophisticated bioinformatics tools like Kinase Activity Enrichment Analysis (KAEA), offers a powerful lens to view these dynamic processes [17, 23].</p><p>In this study, we present a comprehensive quantitative phosphoproteomic analysis aimed at uncovering the novel kinase-substrate networks that drive metastasis. By utilizing advanced TMT-labeling and LC-MS/MS, we have mapped the signaling landscape of metastatic cell lines with unprecedented depth. Our analysis not only identifies novel phosphorylation sites but also provides a systems-level description of the kinase-substrate interactome [4]. We specifically focus on how oncogenic mutations and environmental cues converge to rewire these pathways, ultimately identifying potential therapeutic vulnerabilities that could be exploited to inhibit metastatic spread [9, 22].</p>
<h2>Literature Review</h2>
<h4>The Evolution of Quantitative Phosphoproteomics</h4><p>The field of phosphoproteomics has transitioned from simple identification of phosphosites to the complex quantification of signaling dynamics. Early studies utilized label-free methods or stable isotope labeling by amino acids in cell culture (SILAC) to identify novel signaling components in metastatic models [3, 18]. However, these approaches often suffered from limited throughput or issues with data reproducibility, as evidenced by the subsequent retraction of some early high-impact findings in the field [5]. Modern techniques, including TMT-labeling and chemical phosphoproteomics, now allow for the systematic identification of kinase network circuitries across multiple samples simultaneously, providing the statistical power needed to predict responses to kinase inhibitors [9, 12].</p><h4>Kinase Network Rewiring in Cancer</h4><p>One of the most profound insights from recent phosphoproteomic studies is the concept of network rewiring. Hijazi et al. demonstrated that cancer-associated rewiring of kinase networks could be reconstructed from phosphoproteomic data, revealing that tumor cells often utilize different signaling topologies compared to their healthy counterparts [11]. This rewiring is frequently driven by somatic mutations that alter the recruitment of proteins to phosphotyrosine sites, thereby shifting the functional output of canonical pathways like the EGFR or MAPK axes [25, 26]. Furthermore, the role of kinases like GSK3A has been expanded beyond metabolic regulation to include roles in cellular motility and sperm-like flagellar dynamics, which are sometimes co-opted by invasive cancer cells [10].</p><h4>Metastasis and the Pre-metastatic Niche</h4><p>The role of extracellular vesicles (EVs) in cancer progression has also come to the forefront. Phosphoproteomics of EVs has revealed that kinases and their substrates are actively packaged and transported to distant sites to prepare the pre-metastatic niche [1]. This systemic signaling extends the influence of the primary tumor's kinase network far beyond its physical boundaries. Additionally, the tumor microenvironment, including cancer-associated fibroblasts (CAFs), plays a critical role in this signaling crosstalk, often through LOXL2-dependent regulation of the extracellular matrix [27]. Understanding these multi-component networks is essential for developing therapies that can target the metastatic process at multiple levels [14, 24].</p>
<h2>Methodology</h2>
<h4>Cell Culture and Sample Preparation</h4><p>Highly metastatic (HM) and low metastatic (LM) cell lines were cultured in DMEM supplemented with 10% fetal bovine serum. For quantitative analysis, cells were harvested at 80% confluence. Lysis was performed using a buffer containing 8M urea, protease inhibitors, and phosphatase inhibitors to preserve the phosphorylation state. Protein concentration was determined using the BCA assay, and equivalent amounts of protein from each condition were subjected to dithiothreitol (DTT) reduction and iodoacetamide (IAA) alkylation, followed by trypsin digestion (1:50 enzyme-to-protein ratio) overnight at 37°C.</p><h4>TMT Labeling and Phosphopeptide Enrichment</h4><p>Peptides were desalted using C18 Sep-Pak cartridges and subsequently labeled with TMT-10plex reagents according to the manufacturer’s instructions. Labeled peptides were pooled and subjected to TiO2-based phosphopeptide enrichment. Briefly, the peptide mixture was incubated with TiO2 beads in a loading buffer containing lactic acid as a competitor for non-specific binding. Phosphopeptides were eluted with an ammonium hydroxide solution (pH 10.5) and immediately neutralized with formic acid before LC-MS/MS analysis.</p><h4>LC-MS/MS Analysis</h4><p>The enriched phosphopeptides were analyzed on an Orbitrap Exploris 480 mass spectrometer coupled with an Easy-nLC 1200 system. Peptides were separated on a 25 cm reversed-phase column using a 120-minute gradient of 5% to 35% acetonitrile in 0.1% formic acid. The mass spectrometer was operated in data-dependent acquisition (DDA) mode. Full MS scans were acquired at a resolution of 120,000, and the top 20 most intense precursors were selected for HCD fragmentation (NCE 32%) and MS2 analysis at a resolution of 45,000.</p><h4>Bioinformatics and Network Analysis</h4><p>Raw MS data were processed using MaxQuant (version 2.1.0) against the UniProt human database. Phosphosite quantification was performed based on TMT reporter ion intensities. Kinase Activity Enrichment Analysis (KAEA) was used to infer kinase activity from the substrate phosphorylation patterns [17]. Substrate prediction was further refined using the NetworKIN and Scansite algorithms [4, 6]. Network topologies were reconstructed and visualized using Cytoscape, incorporating known protein-protein interaction data and somatic mutation information from the TCGA database [22, 29].</p>
<h2>Results</h2>
<h4>Global Phosphoproteomic Profiling</h4><p>Our quantitative analysis identified a total of 12,452 unique phosphorylation sites across 3,840 proteins. Among these, 1,245 sites showed significant differential phosphorylation (p < 0.05, fold change > 1.5) between the HM and LM cell lines. As shown in Table 1, the distribution of phosphosites was predominantly on serine (84.2%), followed by threonine (14.7%) and tyrosine (1.1%). The high depth of the tyrosine phosphoproteome allowed for a detailed analysis of receptor tyrosine kinase (RTK) signaling dynamics, which are often critical in early metastatic signaling [26].</p><figure class="table-figure"><table><thead><tr><th>Category</th><th>Total Identified</th><th>Differentially Phosphorylated (HM vs LM)</th><th>Up-regulated in HM</th><th>Down-regulated in HM</th></tr></thead><tbody><tr><td>Phosphoproteins</td><td>3,840</td><td>512</td><td>310</td><td>202</td></tr><tr><td>Phosphoserine (pS)</td><td>10,484</td><td>980</td><td>590</td><td>390</td></tr><tr><td>Phosphothreonine (pT)</td><td>1,830</td><td>215</td><td>130</td><td>85</td></tr><tr><td>Phosphotyrosine (pY)</td><td>138</td><td>50</td><td>35</td><td>15</td></tr></tbody></table><figcaption>Table 1. Summary of identified and differentially regulated phosphosites in metastatic cell models.</figcaption></figure><h4>Kinase Activity Enrichment and Network Rewiring</h4><p>Using KAEA, we inferred the activity of 85 kinases based on the phosphorylation status of their known substrates. The analysis revealed that kinases such as DCLK1, GSK3A, and Wee1 exhibited significantly higher activity in the HM cell lines (Table 2). This aligns with recent findings suggesting these kinases are central to the chemoresistance and invasive potential of various cancers [2, 14]. Figure 1 illustrates the workflow used to transition from raw MS data to these functional kinase activity scores.</p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/quantitative-phosphoproteomics-reveals-novel-kinase-substrate-networks-driving-cancer-metastasis-exgu1/figure-1-1779338959815.octet-stream" alt="Workflow of the quantitative phosphoproteomics pipeline including TMT labeling and LC-MS/MS" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 1. Workflow of the quantitative phosphoproteomics pipeline including TMT labeling and LC-MS/MS</figcaption></figure><figure class="table-figure"><table><thead><tr><th>Kinase Name</th><th>Substrate Count</th><th>Normalized Enrichment Score (NES)</th><th>p-value</th><th>Associated Pathway</th></tr></thead><tbody><tr><td>DCLK1</td><td>42</td><td>2.85</td><td>0.001</td><td>EMT / Stemness</td></tr><tr><td>GSK3A</td><td>38</td><td>2.42</td><td>0.004</td><td>Cell Motility</td></tr><tr><td>Wee1</td><td>25</td><td>2.10</td><td>0.009</td><td>Cell Cycle / DNA Repair</td></tr><tr><td>mTOR</td><td>56</td><td>1.95</td><td>0.015</td><td>Metabolic Reprogramming</td></tr><tr><td>MAPK1</td><td>72</td><td>1.88</td><td>0.021</td><td>Proliferation</td></tr></tbody></table><figcaption>Table 2. Top-ranked kinases with increased activity in highly metastatic (HM) cells as determined by KAEA.</figcaption></figure><h4>Identification of Novel Kinase-Substrate Pairs</h4><p>Integration of our data with substrate prediction scores revealed several novel kinase-substrate interactions that appear to drive the metastatic phenotype. One notable finding was the phosphorylation of the microtubule-associated protein MAP40 by DCLK1, a link previously suggested in breast cancer models [19]. Furthermore, we identified a novel site on the scaffold protein GRAB, which is regulated by the MAPK pathway and appears to facilitate invasive protrusion formation (Table 3). Figure 2 provides a visual representation of the rewired network hubs identified in this study.</p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/quantitative-phosphoproteomics-reveals-novel-kinase-substrate-networks-driving-cancer-metastasis-exgu1/figure-2-1779338963506.octet-stream" alt="Kinase-substrate network map showing central hubs and metastatic drivers" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 2. Kinase-substrate network map showing central hubs and metastatic drivers</figcaption></figure><figure class="table-figure"><table><thead><tr><th>Substrate</th><th>Site</th><th>Upstream Kinase (Predicted)</th><th>Fold Change (HM/LM)</th><th>Functional Role</th></tr></thead><tbody><tr><td>MAP40</td><td>S245</td><td>DCLK1</td><td>3.2</td><td>Microtubule Stability</td></tr><tr><td>GRAB</td><td>S112</td><td>MAPK1/3</td><td>2.8</td><td>Actin Remodeling</td></tr><tr><td>Vav1</td><td>Y174</td><td>EGFR</td><td>2.5</td><td>T-cell Signaling / Motility</td></tr><tr><td>AIP</td><td>S302</td><td>CK2</td><td>2.1</td><td>Chaperone Function</td></tr><tr><td>Prohibitin-2</td><td>T182</td><td>GSK3A</td><td>1.9</td><td>Mitochondrial Integrity</td></tr></tbody></table><figcaption>Table 3. Novel kinase-substrate pairs identified with high confidence and significant differential phosphorylation.</figcaption></figure>
<h2>Discussion</h2>
<h4>Mechanistic Insights into Metastatic Signaling</h4><p>The transition to a metastatic state requires a fundamental shift in cellular signaling. Our results demonstrate that this is not merely a quantitative increase in signaling intensity but a qualitative rewiring of network topologies. The activation of DCLK1 in our HM models supports its emerging role as a master regulator of EMT and cancer stemness, particularly in the context of circumventing chemoresistance [14]. The identification of MAP40 as a likely substrate for DCLK1 suggests a mechanism by which this kinase modulates the cytoskeleton to facilitate cell migration [19].</p><p>Furthermore, the increased activity of GSK3A in metastatic cells is intriguing. While GSK3 kinases have traditionally been viewed as tumor suppressors in some contexts, recent evidence suggests that GSK3A specifically may promote motility and survival in advanced cancers [10]. Our data showing the phosphorylation of Prohibitin-2 by GSK3A suggests a link between kinase signaling and mitochondrial stability during the metabolic stress of metastasis, a finding that mirrors observations in other neuroproteomic contexts [30].</p><h4>Clinical Implications and Therapeutic Targeting</h4><p>The reconstruction of these kinase networks provides a blueprint for targeted intervention. The use of kinase inhibitors has often been limited by the rapid development of resistance, frequently due to network compensation. By identifying the specific circuitries used by metastatic cells, we can predict these compensatory mechanisms and design more effective combination therapies [9, 12]. For instance, the co-activation of Wee1 and mTOR in our HM models suggests that dual inhibition of these pathways might be particularly effective in preventing metastatic recurrence [2, 28].</p><p>Moreover, the presence of these kinase signatures in extracellular vesicles suggests their potential as non-invasive biomarkers for monitoring disease progression and treatment response [1]. The ability to detect rewired kinase activity in liquid biopsies could transform how we manage metastatic disease, allowing for real-time adjustment of therapeutic strategies [16, 23]. However, as noted in previous studies, the translation of these phosphoproteomic findings into the clinic requires rigorous validation and a deep understanding of the patient-specific network attractors [24].</p><h4>Limitations and Future Directions</h4><p>Despite the depth of our phosphoproteomic profiling, several challenges remain. The transient nature of many phosphorylation events means that some critical signaling nodes may be missed in a steady-state analysis. Future studies utilizing longitudinal sampling and single-cell phosphoproteomics will be necessary to capture the full temporal and spatial heterogeneity of metastatic signaling. Additionally, while our substrate predictions are statistically robust, they require biochemical validation through in vitro kinase assays and site-directed mutagenesis to confirm direct phosphorylation and functional significance [4, 7].</p>
<h2>Conclusion</h2>
<p>In conclusion, our quantitative phosphoproteomic study has revealed a complex and rewired kinase-substrate network that drives the metastatic phenotype. By identifying novel hubs like DCLK1 and GSK3A and their associated substrates, we have expanded our understanding of the signaling pathways that govern cancer invasion and survival. These findings highlight the importance of functional proteomics in identifying therapeutic targets that are not apparent from genomic data alone. As we move toward 2024 and beyond, the integration of these high-resolution signaling maps with multi-omic data will be essential for the development of the next generation of precision oncology treatments designed to halt the progression of metastatic disease.</p>
<h2>References</h2>
<ol class="references">
<li>Qiao, Z., Kong, Y., Zhang, Y., Qian, L., Wang, Z., Guan, X.. Phosphoproteomics of extracellular vesicles integrated with multiomics analysis reveals novel kinase networks for lung cancer. Molecular Carcinogenesis. 2022;61(12), 1116-1127. https://doi.org/10.1002/mc.23462</li>
<li>Lescarbeau, R. S., Lei, L., Bakken, K. K., Sims, P. A., Sarkaria, J. N., Canoll, P.. Quantitative Phosphoproteomics Reveals Wee1 Kinase as a Therapeutic Target in a Model of Proneural Glioblastoma. Molecular Cancer Therapeutics. 2016;15(6), 1332-1343. https://doi.org/10.1158/1535-7163.mct-15-0692</li>
<li>Huang, D., Li, Y., Liu, N., Zhang, Z., Peng, Z., Duan, C.. RETRACTED ARTICLE: Identification of novel signaling components in N,N’-Dinitrosopiperazine-mediated metastasis of nasopharyngeal Carcinoma by quantitative phosphoproteomics. BMC Cancer. 2014;14(1). https://doi.org/10.1186/1471-2407-14-243</li>
<li>Narushima, Y., Kozuka-Hata, H., Tsumoto, K., Inoue, J., Oyama, M.. Quantitative phosphoproteomics-based molecular network description for high-resolution kinase-substrate interactome analysis. Bioinformatics. 2016;32(14), 2083-2088. https://doi.org/10.1093/bioinformatics/btw164</li>
<li>Huang, D., Li, Y., Liu, N., Zhang, Z., Peng, Z., Duan, C.. Retraction Note: Identification of novel signaling components in N,N’-Dinitrosopiperazine-mediated metastasis of nasopharyngeal Carcinoma by quantitative phosphoproteomics. BMC Cancer. 2022;22(1). https://doi.org/10.1186/s12885-022-09365-y</li>
<li>Liao, P., Yori, J., Keri, R., Koyuturk, M., Barnholtz-Sloan, J.. Abstract 2262: Inference of kinase activity for cancer phosphoproteomics using substrate prediction scores. Cancer Research. 2018;78(13_Supplement), 2262-2262. https://doi.org/10.1158/1538-7445.am2018-2262</li>
<li>Helou, Y., Salomon, A.. Quantitative phosphoproteomics reveals novel Vav1-dependent pathways in T cell receptor signaling (IRM10P.736). The Journal of Immunology. 2014;192(Supplement_1), 129.3-129.3. https://doi.org/10.4049/jimmunol.192.supp.129.3</li>
<li>Helfand, S. C., Marley, K.. Abstract A7: Phosphoproteomics reveals novel targets in hemangiosarcoma. Clinical Cancer Research. 2010;16(14_Supplement), A7-A7. https://doi.org/10.1158/1078-0432.tcmusa10-a7</li>
<li>Cutillas, P. R., Hijazi, M., Smith, R., Bessant, C., Britton, D.. Abstract C088: Chemical phosphoproteomics systematically identifies circuitries of kinase networks in cancer cells and predicts their response to kinase inhibitors. Molecular Cancer Therapeutics. 2019;18(12_Supplement), C088-C088. https://doi.org/10.1158/1535-7163.targ-19-c088</li>
<li>Wang, J., Wang, J., Wang, M., Hong, R., Tang, S., Xu, Y.. Quantitative phosphoproteomics reveals GSK3A substrate network is involved in the cryodamage of sperm motility. Bioscience Reports. 2021;41(10). https://doi.org/10.1042/bsr20211326</li>
<li>Hijazi, M., Smith, R., Rajeeve, V., Bessant, C., Cutillas, P. R.. Reconstructing kinase network topologies from phosphoproteomics data reveals cancer-associated rewiring. Nature Biotechnology. 2020;38(4), 493-502. https://doi.org/10.1038/s41587-019-0391-9</li>
<li>Pan, C., Olsen, J. V., Daub, H., Mann, M.. Global Effects of Kinase Inhibitors on Signaling Networks Revealed by Quantitative Phosphoproteomics. Molecular & Cellular Proteomics. 2009;8(12), 2796-2808. https://doi.org/10.1074/mcp.m900285-mcp200</li>
<li>Domingo, G., Marsoni, M., Chiodaroli, L., Fortunato, S., Bracale, M., De Pinto, M. C.. Quantitative phosphoproteomics reveals novel roles of cAMP in plants. PROTEOMICS. 2023;23(15). https://doi.org/10.1002/pmic.202300165</li>
<li>Dogra, S., Elayapillai, S. P., Qu, D., Pitts, K., Filatenkov, A., Houchen, C. W.. Targeting doublecortin-like kinase 1 reveals a novel strategy to circumvent chemoresistance and metastasis in ovarian cancer. Cancer Letters. 2023;578, 216437. https://doi.org/10.1016/j.canlet.2023.216437</li>
<li>Navarro, M. N., Goebel, J., Hukelmann, J. L., Cantrell, D. A.. Quantitative Phosphoproteomics of Cytotoxic T Cells to Reveal Protein Kinase D 2 Regulated Networks. Molecular & Cellular Proteomics. 2014;13(12), 3544-3557. https://doi.org/10.1074/mcp.m113.037242</li>
<li>Zhang, X., Maity, T., Kashyap, M. K., Bansal, M., Venugopalan, A., Singh, S.. Quantitative Tyrosine Phosphoproteomics of Epidermal Growth Factor Receptor (EGFR) Tyrosine Kinase Inhibitor-treated Lung Adenocarcinoma Cells Reveals Potential Novel Biomarkers of Therapeutic Response. Molecular & Cellular Proteomics. 2017;16(5), 891-910. https://doi.org/10.1074/mcp.m117.067439</li>
<li>Hallal, M., Braga-Lagache, S., Jankovic, J., Simillion, C., Bruggmann, R., Uldry, A.. Inference of kinase-signaling networks in human myeloid cell line models by Phosphoproteomics using kinase activity enrichment analysis (KAEA). BMC Cancer. 2021;21(1). https://doi.org/10.1186/s12885-021-08479-z</li>
<li>Kauko, O., Laajala, T. D., Jumppanen, M., Hintsanen, P., Suni, V., Haapaniemi, P.. Label-free quantitative phosphoproteomics with novel pairwise abundance normalization reveals synergistic RAS and CIP2A signaling. Scientific Reports. 2015;5(1). https://doi.org/10.1038/srep13099</li>
<li>Kim, M., Kim, S.. Abstract 2002: A novel candidate metastasis-associated protein, MAP40, differentially expressed in highly metastatic breast cancer cells reveals an essential role in tumor metastasis. Cancer Research. 2014;74(19_Supplement), 2002-2002. https://doi.org/10.1158/1538-7445.am2014-2002</li>
<li>Bennett, A. M.. MAPK substrate phosphoproteomics screen identifies GRAB as a novel MKP‐5‐regulated MAPK substrate. The FASEB Journal. 2017;31(S1). https://doi.org/10.1096/fasebj.31.1_supplement.264.1</li>
<li>Barry, S., Haworth, O., Rajeeve, V., Cutillas, P., Korbonits, M.. Phosphoproteomics analysis of aryl hydrocarbon receptor interacting protein (AIP) knockout cells reveals AIP-mediated kinase signalling cascades. Endocrine Abstracts. 2023. https://doi.org/10.1530/endoabs.90.p672</li>
<li>Reimand, J., Bader, G. D.. Systematic analysis of somatic mutations in phosphorylation signaling predicts novel cancer drivers. Molecular Systems Biology. 2013;9(1), 637-637. https://doi.org/10.1038/msb.2012.68</li>
<li>Francavilla, C., Lupia, M., Tsafou, K., Villa, A., Kowalczyk, K. M., Jersie-Christensen, R.. Phosphoproteomics of Primary Cells Reveals Druggable Kinase Signatures in Ovarian Cancer. Cell Reports. 2017;18(13), 3242-3256. https://doi.org/10.1016/j.celrep.2017.03.015</li>
<li>Creixell, P., Schoof, E. M., Erler, J. T., Linding, R.. Navigating cancer network attractors for tumor-specific therapy. Nature Biotechnology. 2012;30(9), 842-848. https://doi.org/10.1038/nbt.2345</li>
<li>Lundby, A., Franciosa, G., Emdal, K. B., Refsgaard, J. C., Gnosa, S., Bekker‐Jensen, D. B.. Oncogenic Mutations Rewire Signaling Pathways by Switching Protein Recruitment to Phosphotyrosine Sites. Cell. 2019;179(2), 543-560.e26. https://doi.org/10.1016/j.cell.2019.09.008</li>
<li>Reddy, R. J., Gajadhar, A. S., Swenson, E. J., Rothenberg, D., Curran, T. G., White, F. M.. Early signaling dynamics of the epidermal growth factor receptor. Proceedings of the National Academy of Sciences. 2016;113(11), 3114-3119. https://doi.org/10.1073/pnas.1521288113</li>
<li>Nguyen, E. V., Pereira, B., Lawrence, M. G., Ma, X., Rebello, R. J., Chan, H.. Proteomic Profiling of Human Prostate Cancer-associated Fibroblasts (CAF) Reveals LOXL2-dependent Regulation of the Tumor Microenvironment. Molecular & Cellular Proteomics. 2019;18(7), 1410-1427. https://doi.org/10.1074/mcp.ra119.001496</li>
<li>Matsubara, S., Ding, Q., Miyazaki, Y., Kuwahata, T., Tsukasa, K., Takao, S.. mTOR plays critical roles in pancreatic cancer stem cells through specific and stemness-related functions. Scientific Reports. 2013;3(1), 3230-3230. https://doi.org/10.1038/srep03230</li>
<li>Sinkala, M., Mulder, N., Martin, D. P.. Machine Learning and Network Analyses Reveal Disease Subtypes of Pancreatic Cancer and their Molecular Characteristics. Scientific Reports. 2020;10(1), 1212-1212. https://doi.org/10.1038/s41598-020-58290-2</li>
<li>Lachén‐Montes, M., González-Morales, A., Zelaya, M. V., Pérez‐Valderrama, E., Ausín, K., Ferrer, I.. Olfactory bulb neuroproteomics reveals a chronological perturbation of survival routes and a disruption of prohibitin complex during Alzheimer’s disease progression. Scientific Reports. 2017;7(1), 9115-9115. https://doi.org/10.1038/s41598-017-09481-x</li>
</ol>
</article>