Working at VU University medical center, Amsterdam
A large part of cancer research aims at biomarker discovery for early cancer detection. These studies are usually performed on high-quality tissue specimens enriched for cancerous cells, whereas the envisioned clinical application usually involves less pure but easily obtainable patient material (i.e. blood, urine, stool, sputum and cervical scrapes/lavages). For these samples, the markers identified in purified tissue samples are likely to be usefull auxiliary information, but may not all be relevant. In this project we will explore empirical Bayes techniques to adapt existing statistical learning algorithms such that these enable use of auxiliary data in an automated and objective manner. These algorithms will be used to identify accurate and robust predictors from molecular profiles of impure patient samples. Specific focus will be on in-house cervical cancer genomics data (microRNA expression and DNA methylation) yielded from self-collected cervico-vaginal specimens (self-samples). Self-sampling can improve cervical cancer screening as it re-attracts up to 1/3rd of the non-responders, in which approximately half of all cancers are found, to the screening program.
Your challenge
The project includes several practical and methodological challenges. It builds upon an existing method that uses auxiliary data (external p-values, genomic annotation, etc) to improve classification accuracy in cancer genomics. All methods will be primarily developed for analysis of the in-house cervical cancer genomics dataset, but will be validated on other cancer genomics data as well.
As a PostDoc, your main tasks and responsibilities are:
Exploring various types of auxiliary data relevant for (cervical) cancer;
Improving variable selection;
Extending it to various types of classifiers and non-binary endpoints such as survival.
We are looking for a candidate who brings the following expertise:
A PhD in (bio)statistics, machine learning or a related field;
Profound knowledge of prediction methods;
Good programming skills (in particular in R);
Fluent in English (scientific writing and speaking);
Able to communicate well with both statisticsians and non-statisticians.
Conditions
Salary scale 10: 2556 tot 4070 euro gross when employed full-time (depending on qualifications and experience).
Salary will be according to the CAO of the Dutch university medical centers. Apart from a good salary, this includes a set 8.3% end-of-year bonus and 8% holiday pay. For more information on all fringe benefits, please visit:www.werkenbijvumc.nl/vumc/arbeidsvoorwaarden/ (Dutch website). We offer a one-year contract in the Department of Epidemiology & Biostatistics of the VU University Medical Center, Amsterdam. This project is financed by the research institute VUmc-CCA. For more information see: www.vumc.nl/afdelingen/cca-site
Interested?
Would you like more information about the postion, please contact dr. Saskia Wilting (biologist) via +31204448584 or Prof. Dr. Mark van de Wiel (statistician) via +31204445405.
Applications
Please apply here: http://www.werkenbijvumc.nl/vacatures/D5.2015.00020RF/?PostDoc+%27Improving+classification+of+molecular+cancer+profiles%27
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|