Héctor Climente

Héctor Climente

Postdoctoral researcher in machine learning & human genetics



I am a postdoctoral researcher the High-Dimensional Statistical Modeling Team, led by Makoto Yamada.

I am interested in bridging the gap between advancements in computer science, most prominently machine learning, and research in biology. In particular, but not exclusively, I study the discordance between the genetic causes of a disease, and the small number of genetic factors we can actually associate with it (aka missing heritability). Complex diseases like familial breast cancer or Alzheimer’s disease are studied through genome-wide association studies. I develop procedures that complement data from these studies with prior knowledge of the disease, with the goal of boosting discovery and selecting more robust features. Ultimately, my objective is to recover interpretable markers of disease which are useful for diagnosis, treatment, and to better understand the biology of the disease.


  • Machine learning
  • Genetics & heritability
  • Biological networks
  • Multidisciplinar research


  • PhD in Bioinformatics, 2020

    Paris Sciences & Lettres

  • MSc in Bioinformatics for Health Sciences, 2014

    Pompeu Fabra University

  • BSc in Biochemistry, 2013

    Autonomous University of Barcelona

  • BSc in Biotechnology, 2012

    Autonomous University of Barcelona



Postdoctoral researcher


May 2020 – Present Kyoto
I am interested in the application of optimal transport to problems in biology, in particular multi-omics data integration.

PhD Student


Oct 2016 – Apr 2020 Paris

CBIO is a joint research team between Mines ParisTech and Institut Curie, devoted to applying machine learning to diagnosis and treatment of cancer. I was supervised by Chloé-Agathe Azencott.

In my thesis, Network-guided genome-wide association studies, I explore existing machine learning algorithms, and develop novel procedures, to better understand the genetics of cancer susceptibility. All those methods leverage on prior biological knowledge, modeled as a graph, to boost discovery.


Head of Biocomputing


Sep 2014 – May 2015 Barcelona
Anaxomics combines systems biology and machine learning to model human cells on which to perform in silico experiments. I was responsible for the development and maintenance of pipelines for the statistical treatment of omics data, whose results were to be included in the models.

Research assistant


Dec 2013 – Aug 2016 Barcelona
In a large-scale study supervised by Eduardo Eyras, we examined the involvement of alternative splicing on 11 different tumor types.




General purpose, dockerized pipelines for GWAS.


Network-based GWAS.


Non-linear feature selection.


Detect differential alternative splicing.