Postdoc From genetics to immune response: A neural network approach
||Voltijd (> 36 uur)
||donderdag 14 september 2017
||donderdag 12 oktober 2017
The objective of the project is to fill gaps in our understanding of the immune system regulation in the context of the host genome, metabolome and microbiome. The task of the PostDoc is to design a computer-simulated artificial neural network model that is complex enough to catch the richness of the immune response. The validated model will be a powerful tool to predict human immune response based on genetic data.
This is a joint program of Radboud University Medical Center and the University of Twente in the framework of the TURBO Grant scheme.
While the cost for sequencing a human genome is only ~$1,000, predicting functionality/phenotype from a DNA sequence, however, is still challenging. Here, we propose a neural network approach to predict human immune response based on genetic data. We have recently initiated the Human Functional Genomics Project (HFGP), which aims to understand the genetic and non-genetic factors influencing human immune responses (Netea et al., Nature Med, 2016). The data comprise the cytokine response of three cohorts of 500 (500FG cohort), 200 (200FG) and 300 (300RO) individuals of Western-European and Eastern-European ancestry, linked to genetic, microbial and environmental factors.
The PostDoc will design a computer-simulated neural network that is complex enough to catch the richness of the cytokine response. After training in 500FG dataset, the model will be validated in the 200FG data set and in a population of a different genetic ancestry (300RO). The validated model will be used to predict the cytokine response in individuals from whom only genetic data are available, which would have important applications in patients with immune-mediated disorders. The ultimate goal, beyond the scope of this computational project, is to realize an efficient hardware realization of the neural network based on the proof-of-concept of the Van der Wiel group (Nature Nanotechnology, 2015). The PostDoc will collaborate closely with other PostDocs and PhD students in an interdisciplinary team of computer scientists, experimental and theoretical physicists, immunologists and mathematicians.
||Starter (0 - 3 jaar)
We are looking for a PostDoc with a PhD degree in computer science and/or mathematics with a strong background in artificial neural networks and machine learning. Experience with developing neural network models for big data search is considered advantageous. Our favored candidates are creative and like to push boundaries. We expect the candidate to have excellent command of the English language as well as professional communication and presentation skills.
Candidates are invited to upload their application at the UT web address, including a short motivation letter, references, and CV to the application button below.
Al ruim dertien jaar publiceert AcademicTransfer alle wetenschappelijke vacatures van de non-profit onderzoeksinstellingen, universiteiten en universitair medische centra in Nederland. De laatste jaren wordt dit complete aanbod aangevuld met vacatures uit het bedrijfsleven waar men ook naarstig op zoek is naar onderzoekstalent.
» Meer over AcademicTransfer