Galicia node consists of 41 members leaded by Carmen Cadarso Javier Roca, José A. Vilar and Francisco Gude doing research in areas such as Generalized Additive Models (GAM) Inference, Extensions to the GAM model, Multi-state Additive models (MSM) in survival analysis


GALICIA NODE

 

1. National Project

Project Title: GENERALIZED ADDITIVE MODELS for ASSOCIATION, PREDICTION and CLASSIFICATION studies. APPLICATIONS to MEDICINE and BIOLOGY.

Project Code: MTM2008-01603.

Main Researcher: Carmen María Cadarso Suárez.

 

2. Researchers that constitute the node

33 researchers from both university and health areas constitute the node. Most of them are active members of the national project MTM2008-01603 that also counts with the participation of researchers from projects MTM2008-03129, MTM2008-00166 and MTM2008-03010. It is a multidiciplinary team including biostatisticians, doctors, biologists, physicists, economists and an IT technician.


Leading Researcher (IR)

Carmen María Cadarso Suárez

  • Researcher responsIble for the Instituto de Santiago Sanitary Research Biostatistics Group (IDIS), recently acredited by the Health Institute Carlos III.
  • Recently admitted as researcher for the Centre for Research in Molecular Medicine and Chronical Diseases (CIMUS) after a binding external evaluation by  high level biomedical scientific commissions. CIMUS is a singular centre for research within the Santiago de Compostela university´s campus of  nternational excellence “Campus Vida“.
  • Researcher of the Belgian European network IAPNetwork P6/03. Statistical Analysis of Association and Dependence in Complex Data.
  • Main Researcher of regional and national projects in both Statistics and Medicine related topics.

 

Researchers from National Universities

Xosé Luis Otero Cepeda

Pablo García Tahoces

Mónica López Ratón

Isabel Martínez Silva

Anaderli Torres Ortiz

Vicente Lustres Pérez

Javier Roca Pardiñas (responsible for node´s Universidad de Vigo section)

Jacobo de Uña Álvarez

Mar Rodríguez  Girondo

José Antonio Vilar Fernández (responsible for node´s Universidade da Coruña section)

María Amalia Jácome Pumar

María Graciela Estévez Pérez

María Ausencia Tome Martínez de Rituerto

Rosa Crujeiras Casais

María del Carmen Carollo Limeres

María José Lombardía Cortiña

 

Researchers from International Universities

Geert Molenberghs

Christel Faes

Thomas Kneib

Luis Meira Machado

Bruno de Sousa

Inês Pereira Silva Cunha Sousa

Luzia Gonçalves

Carlos Daniel Paulino

Giovani Loiola da Silva

 

Researchers from Biomedical Institutions

Francisco Gude Sampedro (responsible for node´s CHUS section)

Pilar Gayoso Diz

María Xosé Rodríguez Álvarez

Arturo González Quintela

Francisco Reyes Santías

Sonia Pértega Díaz

María Teresa Seoane Pillado

Xurxo Hervada Vidal

María Isolina Santiago Pérez

Miguel Ángel Rodríguez Muíños

Alberto Malvar Pintos

Salvador Pita Fernández


Colaboradores científicos

Matías Hisgen

María Pazos Pata


3. Main lines of research

Generalized Additive Models (GAM) represent a flexible tool in multivariate regression studies. In this kind of models the researcher does not need to assume parametric form  for the effects of the covariates in the response. Its versatility allows for several extensions: a) GAM with interactions and unknown link function b) Multistate survival models c)  Multivariate Response Models (VGAM). d) Conditioned ROC curves (on covariates). Amongst these models´ immediate application areas it is worth mention Medicine (Clinical, Epidemiology, Neuroscience and Forensic Medicine and also Biology. In particular, they are essential for Clinical diagnosis and are becoming more and more useful in Epidemiology.

 

Amongst this node´s main lines of research

 

  • Inference in Generalized Additive Models (GAM).
  • Association studies through GAM models.
  • Prediction studies using GAM models.
  • Clasiffication studies through GAM: ROC conditioned on covariates. Optimal combinations of several diagnostic tests. Quantile regression.
  • GAM extensions: Vector GAM (VGAM). Multistate Additive Models (MSM) for Survival.
  • Software development.

 

Apart from the methodological objectives mentioned above and due to the fact that many of the statistical projects developed have appeared as methodological challenges  born from biomedical research projects, it is believed that there is a need to transfer this knowledge to society and those areas of application of biostatistics. This leads to the main objectives:


(a) Application of the proposed methodology to real data from Medicine and Biostatistics areas:

(a1) Measures of association in Clinical Medicine and Epidemiology: non parametric estimation of continuous varuablese effect measurement  curves  like Odds-Ratio (OR), Relative Risk (RR) and Hazard Ratio (HR).

(a2) Neuronal firing rate. Measurement of neuronal time sinchrony. Neuronal population analysis.

(a3) Prediction of Post-Mortem interval in Forensic Medicine through GAM models, SVM (Support Vector Machines) and Neural Networks.

(a4) ROC regression analysis in clinical diagnosis. Statistical evaluation of CAD systems (Computer-Aided Diagnosis) for breast cancer.

(a5) Growth curves in Medicine and Marine Biology. Smooth quantile regression.

(a6) Marine invertebrates reproductive cycle modelling (GAM) applied to marine resources management.

(a7) Modelling of complex survival processes in HIV/AIDS and cancer through flexible MSMs.

(a8) Diseasse mapping.

(a9) Modelling of spatial patterns in biodiversity.


(b) Development of friendly software for those professionals in the medical and biological sciences interested in the practical application of the proposed statistical methodology.


4.  Institutions and groups with whom the node collaborates

  • Xunta de Galicia´s Public Health Directorate (DXSP).
  • Galician Endocrinology and Nutrition Foundation.
  • Marine Research Institute (CSIC - Vigo).
  • Galician Netwok of Biotechnology and Aquaculture (ReGABA).
  • Marine Resources and Fisheries Group – Universidad de A Coruña (UDC).
  • Computational Neuroscience Laboratory (LANCON) – Universidad de Santiago de Compostela (USC).
  • Radiologic Image Laboratory – Universidad de Santiago de Compostela (USC).
  • Aquaculture Institute – Universidad de Santiago de Compostela (USC).
  • Grupo de Parasitología Humana y Animal del Instituto de Investigación y Análisis Alimentarias – Universidad de Santiago de Compostela (USC).
  • Research Group GI-1232 (Marine Invertebrates Laboratory)- Universidad de Santiago de Compostela (USC).

 

5. Biostatistics Consulting Services

Biostatistics consulting services are developed in several centres (Spanish Institutions involved in Galicia Node as shown in chart) :

The CHUS Support Unit for Research provides methodological support to the centre´s different hospitals.

The Coruña University Hospital Complex (CHUAC) Support Unit provides methodological support to the centre´s different hospitals.

USC Medical School Biostatistics Unit integrates research in biostatistics with education and knowledge transfer through cooperation in the execution of the resulting documents and collaboration in research projects in Biomedicine and Statistics (see I.P.´s  curriculum vitae).

The Centre for Research Molecular Medicine and Chronical Diseases (CIMUS) where the IP and her team will be moving soon in will be also carrying out transfer activities in order to execute biostatistical research tasks in the biomedical area. This action will further extend to collaborations with members from the Santiago´s Sanitary Research Institute (IDIS).