Experience

Wellcome Sanger Institute, Hinxton, United Kingdom

Informatics Infrastructure Team Lead

Since Feb 2021

Keywords: Project planning. Resource management. Strategic thinking.

I lead the Informatics Infrastructure team of the Tree of Life programme, which oversees the implementation and delivery of the genome assembly pipelines, and provides support for large-scale genome analyses for the Tree of Life faculty teams.

My ambition is to provide the most efficient platform for assembling and analysing genomes at an unprecedented scale. The Tree of Life projects will generate tens of thousands of high-quality genomes over the coming years – more than have ever been sequenced ! It is a challenging and extremely exciting task that will shape the future of biology.

We are the interface between the Tree of Life teams (assembly production and faculty research) and the Sanger’s IT teams, working together with the informatics teams of the other programmes. The work involves a wide range of scientific fields and technologies such as assembly methods, genomics, comparative genomics, cloud computing, large-scale analyses, with a strong emphasis on metadata tracking, quality controls, and event recording.

European Bioinformatics Institute (EMBL-EBI), Hinxton, United Kingdom

I worked in the Comparative Genomics team of the Ensembl genome browser. The team is in charge of comparing the genomes to one another, implementing new methods and algorithms (extending our API and database schema), and applying them on new datasets. Scalability was our main focus as we had to process hundreds of genomes in a limited timescale.

I was also involved in the development of the eHive workflow manager, a system for creating and running workflows on a distributed compute resource. It is now responsible for scheduling and executing in excess of 1,000 CPU years of compute per year in Ensembl.

Principal Developer

Oct 2019 to Jan 2021

Keywords: Development management. Technical leadership. Recruitment. Mentorship. API development. Database design and optimisation. Workflow design and development. User support (data, API, workflows). Team size: 3 people.

In this role, I transferred the knowledge I had accumulated in the past 8 years to the new Project Leader and the developers, whilst helping them and overseeing the development of the software.

We initiated a massive revamp of our compute workflows and data storage strategy, in order to cope with the scale of data that projects such as the Darwin Tree of Life will generate. Our aim was ambitious: provide comparative analyses on tens of thousands of genomes, and more.

I was still maintaining and contributing to the development of the eHive workflow manager, keeping it the most efficient solution for the Ensembl Comparative Genomics workflows.

Project Leader

May 2014 to Sep 2019

Keywords: Project planning and management. Scientific and public communication. Technical leadership. Recruitment. Mentorship. Development management. Reporting. Data-production planning and operational management. API development. Database design and optimisation. Workflow design and development. Data production under tight deadlines. Processing of large datasets. User support (data, API, workflows). Team size: 2-6 people.

I managed the whole Comparative Genomics team of the Ensembl project, incl. the development of the eHive workflow manager. We had to provide technical support to other Ensembl teams who were using our software. During this period, we wrote an extensive user manual for eHive.

Interim Manager

May 2013 to May 2014

Keywords: Development management. Reporting. Technical advisor. Data-production planning and operational management. API development. Database design and optimisation. Workflow design and development. Data production under tight deadlines. Processing of large datasets. User support (data, API, workflows). Team size: 2 people.

I was managing part of the Comparative Genomics team of Ensembl, incl. the development of the eHive workflow manager, and was still carrying on my developer duties. Our work was focused on the reconstruction of phylogenetic trees and gene families, improving and extending the software. I was still giving some Ensembl API workshops.

Software Developer

Jan 2011 to May 2013

Keywords: API development. Database design and optimisation. Workflow design and development. Data production under tight deadlines. Processing of large datasets. User support (data, API, workflows).

I focused on the reconstruction of protein phylogenetic trees, reshaping the API, improving and extending the software. I also gave some Ensembl API workshops.

École normale supérieure, Paris, France

PhD student

Sep 2006 to Dec 2010

Title: Reconstruction of ancestral vertebrate genomes

I have developed a set of new methods to predict the genome structure of ancestral species (all the last common ancestors of any given group of extant species. Here: about 50 vertebrate species) at different levels (number of chromosomes, chromosome content, gene order). We have also set up a database and a genome browser, Genomicus, to make the data available to the community. As we use the data from the Ensembl project, Genomicus is updated every 2 months, after each Ensembl release.

Thesis (in French) and presentation (in English) available online.

Education

PhD	Bioinformatics	Sep 2006 to Dec 2010	École normale supérieure, Paris, France
MSc	Bioinformatics	Sep 2005 to Aug 2006	Évry university, France
MSc	Computer science, Software development, Mathematics	Sep 2003 to Aug 2006	ENSIIE, Évry, France
BSc	Mathematics	Sep 2004 to Jun 2005	Paris Diderot university, France
Classes préparatoires	Mathematics, Physics, Computer science	Sep 2001 to Jun 2003	Lycée Louis-le-Grand, Paris, France

Projects

2019 Google Summer of Code Mentor

May 2019 to Sep 2019

Title: Using Deep Learning techniques to enhance orthology calls

This project was funded by the 2019 edition of the Google Summer of Code program. Harshit Gupta has been selected to develop a machine-learning algorithm to predict orthologies in the Ensembl Genomes Browser organization under the supervision of myself and Mateus Patricio.

We developed a machine-learning algorithm to predict orthologies, using TensorFlow, directly from sequence data -without using any phylogenetics methods-. The method achieved high accuracy (>90% in most settings) and we are now developing a plan to use it in production in Ensembl.

Project URL: GitHub

2016 Google Summer of Code Mentor

May 2016 to Sep 2016

Title: Graphical editor of XML files

This project was funded by the 2016 edition of the Google Summer of Code program. Anuj Khandelwal has been selected to work on a Graphical workflow editor for eHive using Blockly in the Ensembl Genomes Browser organization under the supervision of myself and Leo Gordon.

eHive is a system used to run computation pipelines in distributed environments. Currently the eHive workflows are configured in a specific file format that requires basic programming skills. This project aimed at removing this drawback by creating a graphical editor for eHive workflows using Google’s Blockly library.

We were envisaging XML as the file format, with a Relax NG specification. The backbone of this graphical editor would be an automated conversion of a Relax NG specification to Blockly blocks and matching rules so that the Blockly diagrams conform to the specification. The graphical interface will have to be able to import existing XML files to visualize them in terms of Blockly blocks, edit them, and export the diagram back to XML.

The project submitted to Google is not specific to eHive and the proposed editor should be able of handling any specifications written using the Relax NG schema.

Project URL: Website, GitHub

EMBL Postdoc Retreat 2013

May 2013

The EMBL postdoc retreat is an official annual EMBL event under the patronage of the EMBL Heads of Units. It promotes scientific exchange among postdocs and provides a platform to address problems relevant to postdocs.

Internship - Web interface for gene annotation

École normale supérieure, Paris, France

Jun 2005 to Aug 2005

I worked on improving the user interface of Exogean, a software for annotating gene structures in eukaryotic genomic DNA.

Contractor - Cost-analysis tool for electronics manufacturers

2005

Conception of a new software to analyse and manage costs for electronics manufacturers (project & product visualisation, resources management).

Personal project - 3D renderer

2002 to 2003

I implemented a 3D renderer from scratch, i.e. without using any libraries such as OpenGL. The program is able to render (project) 3D objects onto a plane as raster images, including colours & transparency, but also lightings and shadings. It’s all written in C++.

I also know about ray-tracing, BSP trees, Bezier curves, B-splines.

Personal project - JPEG encoder/decoder

2001 to 2002

I implemented a JPEG encoder/decoder from scratch, i.e. without using any libraries, reading the documentation of the compression algorithm.

I also wrote a reader / writer for the BMP file format, and a library to manipulate, transform, and apply effects to images.

Certifications

Project Fundamentals Qualification (PFQ) certification

Association for Project Management (APM), 2017

Course on project management. Principles and approaches have since been applied to all Ensembl management layers to plan and manage projects (development and others).

“Policy manager” and “Antivirus Client Security”

F-Secure Corporation, 2004