Proteomic and metabolomic approaches to study diabetic retinopathy
- Autores: Miriam Navarro Sanz
- Directores de la Tesis: María Vinaixa Crevillent (dir. tes.), Oscar Yanes (dir. tes.)
- Lectura: En la Universitat Rovira i Virgili ( España ) en 2018
- Idioma: español
- Tribunal Calificador de la Tesis: J. M. Estanyol (presid.), Noelia Ramírez González (secret.), Mónica Calderón Santiago (voc.)
- Programa de doctorado: Programa de Doctorado en Biomedicina por la Universidad Rovira i Virgili
- Materias:
- Enlaces
- Tesis en acceso abierto en: TDX
- Resumen
The general objective of this doctoral thesis was to develop, analyse and validate new bioinformatic tools for converting raw MS-based metabolomics data into biological knowledge, in order to study alterations in the proteome and metabolome of human retinal pigment epithelium cells exposed to hyperglycemic and/or hypoxic conditions.
To reach this general objective, I have structured my thesis in two blocks:
⁻ Methodological aims: (i) analyse mass spectral databases for LC/MS-based untargeted metabolomics, and (ii) generate and improve the characterization of LC/MS metabolomics data focusing on MS1 and MS2 annotation.
⁻ Biological aims: (iii) detect and analyse changes in protein-protein interaction (PPI) networks by hyperglycemic and/or hypoxic conditions, and (iv) predict and validate metabolite alterations due to hyperglycemic and/or hypoxic conditions integrating protein expression data in metabolic networks.
My contribution to aim (i) was a thorough study of the NIST14 mass spectral database, proving it superior to other available databases due to its larger number of metabolites with spectral data acquired from different adducts and using a wide range of mass spectrometers. I showed the importance of adduct formation for metabolite identification, analysing MS/MS data for different adducts in my lab. This is particularly important because predominant adducts in an ESI spectrum vary from one metabolite to another as well as on the mobile phase used (Vinaixa M et al. TrAC Trends in Analytical Chemistry, 2016).
My contribution to aim (ii) was to evaluate the performance of two new computational tools: CliqueMS and iMet, which were developed in collaboration with SEES lab led by Dr. Roger Guimerà and Marta Sales-Pardo (URV). CliqueMS annotates in-source MS1 data based on a coelution similarity network. I have experimentally proven that CliqueMS correctly identifies and annotates a large number of adducts from pure standards and complex biological samples, leading to more correctly parental ion neutral masses than CAMERA, the most widely-used approach (Senan et al. Bioinformatics, under revision).
iMet facilitates structural annotation of metabolites based on MS2 data not described in mass spectral databases. I simulated a real scenario of metabolites not present in a database by testing iMet using metabolites proposed in the Critical Assessment of Small Molecule Identification (CASMI) challenges from years 2012-2016. In addition, I compared iMet’s performance to other tools such as CFM-ID, MetFrag and MS-Finder. I could demonstrate the potential of iMet for annotating metabolites that are not present in databases, as well as to compare the performance of other methods to assist the structural annotation of known metabolites lacking MS/MS spectra in databases (Aguilar-Mogas et al. Analytical Chemistry, 2017).
On the other hand, the biological applications of my thesis aimed at studying diabetic retinopathy (DR) using metabolomics and proteomics approaches. It is important to highlight that the training and results obtained above was key to improve the metabolome coverage in ARPE-19 cells and vitreous humour samples.
My contribution to aim (iii) was to generate proteomics data and develop a novel approach that integrates PPI, module analysis and protein expression for detecting dysregulated groups of interacting proteins involved in similar biological processes. This work was performed in collaborations with the Structural Bioinformatics and Network Biology group led by Dr. Patrick Aloy at the IRB Barcelona, and with the SEES lab at the URV. Using this new approach, it has been possible to capture slight but consistent protein changes occurring in a protein module which are impossible to detect considering only individual proteins. (Navarro M et al. In preparation, 2018).
My contribution to aim (iv) was to validate a novel proteomics data analysis workflow based on a human genome-scale metabolic network that predicted metabolic alterations in an in vitro model of DR. I have generated and analysed metabolomics data on ARPE-19 cells cultured at low and high glucose concentrations, and normoxic or hypoxic conditions, also fed with 13C-glucose for isotopic label tracking (flux analysis), to validate the predictions made by our novel data analysis workflow. In addition, I also analysed human vitreous humor from DR patients and controls for clinical validation.
