2010-04-07 / 16:32 / dave
Precision Therapeutics, Inc: Job Description
Job Title: Informatics Intern
Precision Therapeutics, Inc. is a diagnostics services company dedicated to providing physicians and patients with actionable clinical information to personalize cancer treatments. We are currently seeking an Informatics Intern.
Duties and Responsibilities
(Include the following: other duties may be assigned)
The goal is to do a thorough investigation on the best way to measure in vitro chemosensitivity generated by ChemoFx assays. Traditionally the two major schools of thought are IC50 and AUC of the dose-response curves. Due to the special features of the curves from ChemoFx assays (e.g. non-monotone, resistant lines do not respond, non-sigmoidal), there are some AUC derivative metrics (such as aAUC, tAUC) that are intended to account for the curve characteristics that are not captured by AUC. However, it has been unclear which measure is the ‘best’, and it is difficult not to pre-specify this in a protocol.
The current thought is that this research involves two components. (1). Mathematically, what is the optimal way to quantify a dose-response curve from in vitro assays? This step involves the combination of mathematical models AND the knowledge of in vitro biology (especially w.r.t. how ChemoFx is setup). Both parametric models (such as logistic regression model, polynomial model; as well as the choice of IC50, IC25, etc) and non-parametric models (e.g. AUC and its derivatives, smoothing spline) need to be considered. (2). Use clinical data that have already demonstrated the ‘link’ between ChemoFx assay and clinical outcome to compare the performance of different metrics. This can be achieved by in-sample cross-validation approaches or by using multiple independent datasets.
- Graduate student with 2 years towards PhD in Statistics
- No prior working experience is needed
- R and/or SAS programming
- Good communication skills and writing skills
- Able to perform statistical simulations
- Knowledge of prediction and cross-validation
- Knowledge of non-linear regression and model fitting, especially the 4-parameter logistic regression model for dose-response curves
- Knowledge of area under the curve (AUC) calculation
If you are seeking a dynamic, challenging atmosphere, that is never boring, with a chance to make a difference and help cancer patients, email your resume to PTIResumes@ptilabs.com with the word “Informatics Intern” in the subject of the email
I didn’t write the posting, but feel free to ask me questions.