- PhD -
Molecular biologist, bioinformatician, data scientist
Current position, from 2015
Faculty of Veterinary Medicine, University of Zagreb - Zagreb, CROATIA
Establish a new laboratory of genomics
Molecular biology is a new research way in veterinary medicine in Croatia. Objective of the project is to finance, build and equip a new laboratory in molecular biology (proteomics and genomcis). I was recruited to this project as an expert in genomics and bioinformatics. I bring my experience and my skills to create experimental protocols in genomics, like DNA and RNA extractions, PCR and qPCR.
Project VetMedZg: ERA Chair FP7 - Upgrading the research performance in molecular medicine at the faculty of veterinary medicine, University of Zagreb
Genomic laboratory before equipment
Operational genomic laboratory
Consultant in genomics
As an experienced researcher in genomics, I train different researchers from the Faculty to add and use genomics in their current research projects (DNA and RNA extraction, nucleic acids handling, PCR and qPCR). I worked notably with projects related with cancer and parasitology in dogs. I offered my support for different projects writing.
Ressource person for bioinformatics and data analysis
I perform statistical and bioinformatics analysis for our laboratory experiments, but also for collaborators, which can be teams at the Faculty but also visitors from different countries (Brazil, UK, Iran).
My aim is to transform raw data into understandable and usable knowledge.
In bioinformatics, I perform Gene Ontology analysis to better characterized complex phenotypes and diseases, and predict potential new markers.
Creation of a new diagnostics test
Answering an ask from the Department of Parasitology, I set up a new test to diagnose presence in blood of Babesia canis canis, a blood parasite in dogs which cause a potentially lethal disease, babesiosis.
I designed a PCR-bases test which can diagnose presence of less than 1% parasitism rate of babesia in 2 hours, faster and more accuratly than microscopic identification and counting (which is he usual method used). Due to its specificity, this test can discriminate 2 closely related subspecies of babesia, Babesia canis canis and Babesia canis vogeli. This is an important feature as the 2 subspecies can be cured with different drugs.
Development of a statistical pipeline tool to quickly analyse thousands of proteomics data generated by LC-MS
Our laboratory LC-MS generate thousands of peptides quantification on many different samples. I encoded a pipeline tool in R langage to manage and analyse such big data. This pipeline import data generated by LC-MS, calculate different descriptive statistics (fold change, mean, standard deviation, etc...) and finally identify peptides with significant changes of expression. The pipeline can be used for every kind of experimental plan, from 2 to an unlimited number of groups. In few seconds, the pipeline run the comlpete analyse of thousands of peptides data and export results into understandable tables.
This pipeline has been used not only for our own researches in the laboratory VetMedZg, but also on our partners data.
During the school year 2016-2017
French School of Zagreb, High School division - Zagreb, CROATIA
Teacher in Life and Earth Sciences - High School
I was contacted by the school to take in charge courses in Life and Earth Sciences for 5 high school students. The school has issues to find a french-speaking teacher for this thematic, and I was pleased to support French community in that way. During 1 school year, I tought to few teenager students. It was a wonderfull experience, as it change completly from my "routine" laboratory work. It also make me think differently, as I usually teach to established researchers. So I adapted myself to teach to teenagers. I organized some practical work in my laboratory to make my students discover what is research in molecular biology. At the end of the year, the only last-year graduation student in the group successfully validated his final exam in Life and Earth Sciences. I was happy to lead him to the success !
Saint-François d'Assise Hospital, CHUQ - Québec, CANADA
Support for project submission - Bibliography and data analysis
My aim for this position was to provide a complete bibliography report on the utilization of antibiotics in primary care, and summarize statistics about such utilization. Using such data, I provided support for the submission of research projects about antibiotics in primary care. I also summarize data about elderly care. Thanks to this job, I discovered closely research in an hospital with thematics and problematics, directly applied to patients.
Center for Reproductive Biology Researches, University Laval - Québec, CANADA
Establish a new methodology to identify genetic markers from genomics data generated by NGS
My post-doctoral aim was to create a analyse pipeline to identify genetic markers (SNP) of complex phenotype, by the application of genomics. The studied phenotype was the expression of a gene encoding the aromatase enzyme, essential component of female fertility, in dairy cattle.
I quantified gene expression of this gene in 100 samples. Then I studied gene expression by a DNA micro-array to determine which genes could be related with a modification of the aromatase gene expression profile. After sequencing bovine genome by Illumina HiSeq on a reference population (N=1000) and alignment, I genotyped the 100 samples on the best candidates genes (from DNA micro-array experiment) to characterize their genome on specific positions (SNP).
To modelize aromatase gene expression profile in function of genotype content, I was inspiered by publications in human forensic science (determination of hair color by genetic markers). Then I calculated a prediction model and identify the best genetic markers related with the complex phetnotype (aromatase gene expression level).
Before the end of my post-doctoral internship, I set up this metholodology to identify genetic markers using genomics data, allowing a genomics laboratory to identify genetic markers. I also wrote protocol for genotyping, as I was the first to perform such experiment in the Research Center.
Identification and validation of new genetic markers (SNP)
With the utlization of the methodology set up, I identified and validated genetic markers related with aromatase gene expression profile. Those SNPs were then validated in bulls, according to their daughters performances.
Some of those markers were completely unknown by the industry. Some were located in known QTL and so allow to precise coordinates of this region.
In those two cases, my work identified new markers for dairy cattle indstry.
Herbivore Research Unit, National Institute for Agricultural Researches (INRA) - Clermont-Ferrand-Theix, FRANCE
Set up a new methodology for muscle protein quantification
At the beginning of my PhD, I needed a better methods (in terms of speed) than Western-Blot to quantity proteins in muscle. So I started to set up a new method to quickly quantify proteins, with at least the same accurac than Western-Blot. After 1 years of tests and improvments, I successfully publish a protocol of proteomics dot-blot. By the utlization of this technique, I was able to quantify proteins in hundred of muscle samples in few weeks, intead in 2 to 3 years as it should have been with Western-Blot. This technique changed completely my PhD results.
Set up fluorescence detection of antibodies in the laboratory
When I started my PhD, the only detection system of antibodies used in the laboratory was chemiluminescence. When I was developing Dot-Blot protocol, it appear that I needed a better detection system in term of sensitivity and reproductibility. I was trained in the fluorescent detection system by a laboratory in human physiology, and bring the protocol back to my laboratory (with some adaptation to the equipment). This protocol is now a routine one in the laboratory.
Set up a new statistical approach for the modelization of complex phenotype from proteomics data
Main aim of my PhD was to validate (or not) the relation of different muscle proteins with beef tenderness. For that, I calculated prediction models of beef tenderness score from proteomics data, using a dedicated methodology designed for this objective.
Validation of proteins as biomarkers and pathway analysis of beef tenderness
Thanks to my PhD results, some proteins were validated as markers of beef tenderness. Going further, I performed bioinformatics analysis to decipher beef tenderness into cellular pathways. Using those analysis, I predicted 3 never-studied (in beef tenderness) proteins as potential new biomakers. Those proteins were validated few months after my PhD for their role in beef tenderness.
2006 - 2007
Laboratory of microbiology and genetics, University Nancy II - Nancy, France
Design a new bacterial plasmid in Streptomyces ambofaciens
Aim of this internship was to design a new plasmid by molecular cloning and transform the soil bacteria Streptomyces ambofaciens with it.
During this internship (in alternance with Master courses), I successfully designed the required plasmid. Due to toxicity of the genetic construction for Escheriachia coli, I needed to transformed directly Streptomyces ambofaciens which is a more time-consuming process. I produced different strains of bacteria for laboratory collection.
I had working experiences in 3 different countries, on 2 continents. I worked with different people with different working philosophy, on different thematics and objectives. All those various and rich experiments forged the researcher I am today.
A Gene Ontology analyse to decipher pathways implied in a dog infectious disease
Diagnostic by qPCR for the detection of Babesia canis canis in dog blood
Part of the developed script for LC-MS big data analyse
Volcano plot of protein fold changes and their associated p-values
Results of HRM genotyping - 3 different genotypes detected on 1 position
An example of DNA micro-array
Identification of significant (or not) SNPs related with studied phenotype
Volcano plot of more than 1000 gene expression data produced by DNA micro-array
First Dot-Blot. This methodology saved 2-3 years of PhD.
Switching from chemiluminescence to fluorescence detection produced better results (sensibility, reproductibility)
Functional interactome generated from my PhD results
Predictive model for beef tenderness