The CellCycler is a minimalist model of a growing tumour which uses ordinary differential equations (ODEs) to simulate cells as they pass through the phases of the cell cycle. The guiding philosophy of the model is that it should only use parameters that can be observed or reasonably well approximated. There is no representation of the complex internal dynamics of each cell; instead the level of analysis is limited to cell state observables such as cell phase, apoptosis, and damage. This approach naturally accounts for a heterogeneous cell population with varying doubling time, and closely captures the dynamics of a growing tumour as it is exposed to treatment (see related post).
The Docetaxel survival app for metastatic castrate resistant prostate cancer (mCRPC) provides a prediction on the survival prognosis of a patient based on the value of two variables which are to be collected just before treatment is initiated. Although the model may be simple it has been tested on both a large clinical trial data-set and also a real world data-set taken from the Chrsities Hospital in Manchester, UK. The model was also tested against the most recent model and found to be superior. Publications related to the model will be made available soon and will be cited.
The Ion-Channel Cardiac Toxicity Estimator app is for the safety scientist who is not keen on large-scale models of the heart but wants similar level of predictivity as those models using a model that is far easier to understand and interpret. The app is based on a publication which has shown that a simple division can produce and maybe improve upon the performance over large differential equations models which contain 100s of parameters. The app allows you to compare the predictions across four different assays including both animal and human. By providing an alternative solution to the mainstream we hope it encourages people to question others around the scale of model used for this prediction task. Related post.
The BayesianOpinionator is a web app for incorporating the effect of prior beliefs when determining the impact of a statistical study. The study data is assumed to be in the form of a comparison between two cases, denoted null and treated. For example in a clinical trial the treated case could correspond to a patient population who are treated with a particular drug, and the null case would be a comparison group that are untreated. A common problem with such studies is that they produce results which appear to be statistically significant, but later turn out to be caused by a fluke. In this case the BayesianOpinionator will help to determine how seriously the results should be taken, by taking prior beiefs and data into account. Related post.
The Lending club profit/loss app provides an investor the probability of gaining a return of a certain amount given the credentials of the loan applicant. The model was developed using data from the lending club and applying survival modelling techniques to the data. The final model simply requires three bits of information, interest rate, term of loan and the last FICO high range value. The output of the model is a profitability curve which provides probability of making a loss as well as a certain amount of profit. Related post.
The RentOrBuyer app provides a level playing field for comparing the costs of buying and renting a house. It works by considering two scenarios. In the Buy scenario, the costs include the initial downpayment, mortgage payments, and monthly maintenance fees (including regular repairs, utilities, property taxes, and accrued expenses for e.g. major renovations). The Rent scenario has exactly the same initial and monthly expenses, however the costs in this case only involve rent and utilities. The initial downpayment is therefore invested, as are any monthly savings compared to the Buy scenario. By showing which of these is worth more, we can see whether it is financially better to buy or rent. Related post.