This opportunity has expired. It is therefore no longer possible to like or apply.
See similar opportunities insteadYour career starts on Magnet.me
Create a profile and receive smart job recommendations based on your liked jobs.
How does your internship at Rijk Zwaan look like?
Hundreds of thousands of scientific articles are currently published every year. These articles bring a wealth of knowledge that is difficult to grasp manually. Obtaining complete information from these articles is an impossible manual task. Various text mining tools have been developed in the past to support this by identifying relevant information from the articles in a (semi-)automatic way.
Our current text mining research tool based on BioBERT architecture (Lee, 2019), allows us to automatically extract bio-entity terms and their relations from research articles. This makes it possible to further build a fully scalable knowledge graph. The goal of this project is to build a knowledge graph using BioBERT, and then predict the indirect gene-trait links by combining information from different scientific articles.
The knowledge graph will be built using graph database solutions and we will explore different graph algorithms to predict the indirect relations between genes and traits (Mark, 2019). Various link prediction algorithms, such as Adamic Adar, Common Neighbors etc. will be benchmarked based on its relevant score prediction. The relevance of such predictions will also be further assessed by our trait experts. This feedback will allow improvement of the prediction algorithm to increase the prediction score, resulting in more relevant predictions. The outcome of this project is to report the benchmarking results of applying different link prediction algorithms.
During this data science internship, you will get opportunity to learn and improve your knowledge about graph database solutions and link prediction algorithms. You will be working with bioinformatics researchers, data scientist and trait experts.
Bibliography
Lee, J. a. (2019). BioBERT: a pre-trained biomedical language representation model for biomedical text mining. Bioinformatics.
Mark Needham, A. E. (2019). Graph Algorithms. O'Reilly Media, Inc.
Your contribution:
You are a curious person with a positive attitude to research activities. Moreover you are a team player with a constant thrive to improve.
In addition, we can find the following on your resume:
What can you expect from us?
Besides an growing international work environment, we are happy to offer you:
Rijk Zwaan: your new employer?
In Fijnaart, West-Brabant, you can find a Research & Development location of Rijk Zwaan.
Rijk Zwaan is an international family company with a people-oriented culture. Rijk Zwaan is right at the start of the food chain. We develop vegetable varieties and sell the seeds produced from them globally.
From our strategic position we contribute to the world food supply. Everything we do is about growth. The growth of natural products that we export around the world. The growth of our company, which currently employs over 3,300 people in more than 30 countries. And above all the growth of our people: we give you every opportunity to use your own initiative and achieve your full potential.
Het primaire doel van Rijk Zwaan is medewerkers een prettige en duurzame baan bieden met goede arbeidsvoorwaarden. In ons familiebedrijf krijgt iedereen de ruimte om te werken als zelfstandig ondernemer. De werksfeer is motiverend en informeel. Een typische Rijk Zwaan collega is betrokken en loyaal en werkt graag samen met anderen.
"Bij digitaal fenotyperen worden de uiterlijke kenmerken van het gewas – het fenotype – beoordeeld door de computer aan de hand van foto’s. Zo weten we snel en eenduidig welke meloen de gewenste eigenschap heeft.
De grootste uitdaging? Zorgen dat de computer elke foto op dezelfde manier beoordeelt."
"Bij Rijk Zwaan ontwikkelden we een innovatieve eigenschap in sla. Door deze eigenschap worden de snijranden van gesneden sla minder snel bruin. Hierdoor blijft de sla minimaal twee dagen langer houdbaar
Change language to: Dutch
This page is optimised for people from the Netherlands. View the version optimised for people from the UK.