Introducing Maspectra

We are very happy to introduce Maspectra® to help you accelerate the modeling & simulation process for drug development.

Maspectra® is designed to be a data analysis platform that excels in performance, usability, and extensibility, specifically catering to the diverse needs of pharmacology and statistics functionalities.

Our development mainly focused on two aspects: a Python-based modeling framework named masmod and a cross-platform software graphical user interface.

masmod framework

masmod is a high-performance pharmacokinetic/pharmacodynamic (PK/PD) modeling & simulation framework with a well-designed Python interface. It provides a set of tools including exploratory data analysis, model definition/estimation/evaluation, and simulation.

The syntax of masmod follows Python's widely recognized style, making it a breeze to learn and use. Its extensibility enables effortlessly incorporating your own models into a customized masmod model library.

Writing and reading a masmod model is as easy as writing and reading math equations. You can also use masmod to simulate from models and visualize the results. If you have experience with the renowned machine learning framework PyTorch, you'll notice the similarities in syntax, making it intuitive and accessible.

Show me a demo!

/models/warfarin_final.py

class DemoPopPKProblem(Module):


    def __init__(self):
        super().__init__()

        # pk parameters
        self.tv_v = theta(8.06, bounds=(0.01, 20))
        self.tv_cl = theta(0.131, bounds=(0.01, 1))
        self.tv_ka = theta(1.325, bounds=(0.01, None))
        self.tv_alag = theta(1.03, bounds=(0.01, 24))
        self.eta_v = omega(0.016744)
        self.eta_cl = omega(0.0662868)
        self.eta_ka = omega(0.452064)
        self.eta_alag = omega(0.302627)
        self.eps_prop_pk = sigma(0.00478)
        self.eps_add_pk = sigma(0.135)

        # Compartments
        self.dose_cmt = compartment(default_dose=True)
        self.central_cmt = compartment(default_obs=True)

        # From Data Input
        self.wt, self.dvid = column(['WT', 'DVID'])

    def pred(self):
        cl = self.tv_cl * (self.wt / 70)**0.75 * exp(self.eta_cl)  # Clearance
        v = self.tv_v * (self.wt / 70) * exp(self.eta_v)  # Volume
        ka = self.tv_ka * exp(self.eta_ka)  # Absorption rate
        alag = self.tv_alag * exp(self.eta_alag)  # Absorption lag time
        k20 = cl / v

        self.dose_cmt.alag = alag  # Alag on dosing compartment
        cp = self.central_cmt.A / v  # Central compartment concentration
        self.dose_cmt.dAdt = -ka * self.dose_cmt.A  # Dose compartment dynamics
        self.central_cmt.dAdt = ka * self.dose_cmt.A - k20 * self.central_cmt.A  # Central compartment dynamics

        y = cp * (1 + self.eps_prop_pk) + self.eps_add_pk  # Additive and proportional error
        return y

If you are interested in masmod, please check out our quickstart for more details.

Maspectra® graphical user interface

Maspectra® is a desktop software for modeling & simulation workflow. It provides a graphical user interface (GUI) that plays the role of an integrated development environment (IDE) for Modeling Information Drug Discovery and Development (MID3).

It is an all-in-one solution for MID3 workflows, including pharmacokinetic/pharmacodynamic modeling, statistical analysis, data processing/visualization, and more. The software is cross-platform, which means it can run on Windows and macOS (Linux support is under development).

We will give more details about the Maspectra® software in future blogs. Keep an eye on us 👀!

This is the first post of this blog. Thanks for reading. We will keep updating this blog to share our progress with you. If you are interested in our project, please follow us on GitHub. Also, we will post some more interesting things here. Stay tuned!

tip

We recommand you to join our wait list to experience Maspectra® first-in-place as soon as it becomes available to public.