After summer break the applied mathematics seminar will continue.

On Thursday, Sep. 17th, Teemu Daniel Laajala  will give a talk starting at 10.15 at seminar room M1 (Quantum, 2nd floor)

Title:  Improving the Design and Analysis of Preclinical Experiments

Abstract:
Preclinical cancer intervention experiments are currently conducted with suboptimal design and appalling reproducibility [1-5]. Here, I will present a matching-based approach for improving the robustness in statistical inference in such longitudinal preclinical experiments. Firstly, using refined non-bipartite matching, optimal subgroups of animals are detected at baseline, and are allocated in a balanced and blinded manner to intervention groups. After conducting the experiments with fully masked intervention labels, the matching information is utilized to increase statistical accuracy in the inference of possible intervention effects. A mixed-effects model for performing timepoint-specific matched inference is presented for modeling the post-intervention longitudinal tumor growth profiles. Lastly, model-based power simulations are proposed to be coupled with the mixed-effects models to estimate sufficient sample sizes for future experiments.

References

1. Begley, C.G. & Ellis, L.M. Drug development: Raise standards for preclinical cancer research. Nature 483: 531-533 (2012).
2. Couzin-Frankel, J. When mice mislead. Science 342: 922-925 (2013).
3. Cressey, D. UK funders demand strong statistics for animal studies. Nature 520: 271-272 (2015).
4. Perrin, S. Preclinical research: make mouse studies work. Nature 507: 423-425 (2014).
5. Macleod, M. Why animal research needs to improve. Nature 477: 511 (2011).

All interested are warmly welcome!

After summer break the applied mathematics seminar will continue.

On Thursday, Sep. 17th, Teemu Daniel Laajala  will give a talk starting at 10.15 at seminar room M1 (Quantum, 2nd floor)

Title:  Improving the Design and Analysis of Preclinical Experiments

Abstract:
Preclinical cancer intervention experiments are currently conducted with suboptimal design and appalling reproducibility [1-5]. Here, I will present a matching-based approach for improving the robustness in statistical inference in such longitudinal preclinical experiments. Firstly, using refined non-bipartite matching, optimal subgroups of animals are detected at baseline, and are allocated in a balanced and blinded manner to intervention groups. After conducting the experiments with fully masked intervention labels, the matching information is utilized to increase statistical accuracy in the inference of possible intervention effects. A mixed-effects model for performing timepoint-specific matched inference is presented for modeling the post-intervention longitudinal tumor growth profiles. Lastly, model-based power simulations are proposed to be coupled with the mixed-effects models to estimate sufficient sample sizes for future experiments.

References

1. Begley, C.G. & Ellis, L.M. Drug development: Raise standards for preclinical cancer research. Nature 483: 531-533 (2012).
2. Couzin-Frankel, J. When mice mislead. Science 342: 922-925 (2013).
3. Cressey, D. UK funders demand strong statistics for animal studies. Nature 520: 271-272 (2015).
4. Perrin, S. Preclinical research: make mouse studies work. Nature 507: 423-425 (2014).
5. Macleod, M. Why animal research needs to improve. Nature 477: 511 (2011).

All interested are warmly welcome!