Disease Course Mapping with Leaspy#
by Igor Koval
Disease Course Mapping with Leaspy
Leaspy is a Python software package that implements the Disease Course Mapping methods. In particular, it is designed for the statistical analysis of longitudinal data, particularly medical data that comes in a form of repeated observations of patients at different time-points. Considering these series of short-term data, the software aims at :
recombining them to reconstruct the long-term spatio-temporal trajectory of progression
positioning each patient observations relatively to the group-average progression
quantifying the impact of different cofactors (gender, genetic mutation, environmental factors, etc) on the progression of the biomarkers
imputing missing values
predicting future observations
simulating virtual patients
Software Python package Open source Tutorials Continuous integration
The software is distributed on Gitlab under the GNU GPLv3 Licence. It has a complete documentation as well as an active community that you can solicit through the use of the dedicated bug & issue tracker. The software operates on Mac and Linux - Windows is also working though no guarantee can be given as no specific development was developed for this platform.
Note
Leaspy originally comes from from LEArning Spatiotemporal Patterns in Python.
Usage#
The package has been written to offer a user-friendly API in order to be used by non-experts. The API essentially includes the following functions
leaspy.fit(...)To estimate the average progression of the biomarkers
leaspy.personalize(...)To estimate the individual parameters that allows to derive the average progression to fit individual Data
leaspy.estimate(...)To impute or predict the values of an individual at any time-point
leaspy.simulate(...)To simulate synthetic patients
leaspy.load(...)&leaspy.save(...)To load and save the entire description of the disease progression, share it or reuse it on another cohort
Tip
New developers are welcome to participate and contribute !
While offering an easy-to-use API, the package has been designed with an internal modular architecture in order to make new developments possible. In particular, new developments are on they way to enlarge the possibilities of types of disease progression, in particular with ordinal data or progressions affected by drugs.
Tutorials#
We have developed few tutorials to better understand the goals of Disease Course Mapping, in particular how it allows to go beyond the current limitations of linear mixed-effects models, while getting familiar with Leaspy.
Tutorial 1: Limitations of linear mixed-effects model
This introduction to longitudinal data progressively unveils the limitations of linear models, beginning by a cross-patient linear regression, going to individual linear regressions, then to a linear mixed-effects model, finishing by the limitations of linear models that Leaspy is able to overcome.
Go to the tutorial 1 hour longitudinal data Linear model Mixed effects model
Tutorial 2: Hello World with Leaspy
This introduction to Leaspy gives a handful overview of Leaspy possibilities along with the user-friendly commands to use it
Go to the tutorial 30 minutes longitudinal data non-linear mixed-effects model
Tutorial 3: Real-case usage
This tutorial is designed for the ones that want to use Leaspy with their own data. We therefore designed this tutorial to present step-by-step the different operations that you might go through : data manipulation, normalization, interpretation of the algorithm convergence, in-depth understanding of the parameters and hyperparameters, etc.
Go to the tutorial 1 hour 30 longitudinal data Missing values longitudinal data
Installation#
The installation procedure is entirely detailed on the dedicated Gitlab repository