[Vp-reproduce-subgroup] MODEL REPRODUCIBILITY, CREDIBILITY AND STANDARDIZATION SUBGROUP

[Rahuman Sheriff]

Dear all,
I am Sheriff, Project Leader at European Bioinformatics Institute (EMBL-EBI), Hinxton, Cambridge, UK.
Jacob invited me to co-lead this subgroup and I gladly accepted it.

Please find below a brief description of our subgroup's focus and key issues we will try to address.

Btw, I am giving a talk on “Reproducibility of models” on January 21  (Thursday) at 3pm Eastern Standard Time

https://iu.zoom.us/meeting/register/tZYqd-2srD8tGtCXDem4Cka08rBz5fDW0EQR


I will talk about our recent work, where we analysed 455 models from published articles and found that half of them could not be reproduced
using the information provided in the manuscript. I will discuss the reasons why models often fail to reproduce and the solution we propose to tackle this during the peer-review process.

Full BioRxiv manuscript here
https://www.biorxiv.org/content/10.1101/2020.08.07.239855v1 <https://www.biorxiv.org/content/10.1101/2020.08.07.239855v1>
This is now accepted in Mol Sys Bio and I will present some more results.


I am Jacob had some initial discussion, we have some excellent plants for our groups activities. Hopefully we can meet soon and discuss them.


Best regards
Sheriff






MODEL REPRODUCIBILITY, CREDIBILITY AND STANDARDIZATION SUBGROUP
MSM WORKING GROUP ON MULTISCALE MODELING OF VIRAL PANDEMICS

Computational biological and medical models attempt to capture knowledge in a way that can be more easily shared than human language. The number of computational medical models are increasing. However, reuse of models is still rare indicating that there is still a gap that needs to be closed for computational models to become useful. Even published models are rarely reusable out of the box for multiple reasons. Many times the models just cannot be reproduced. 
One major reason is the lack of adoption of community standards by modellers despite many efforts. Model codes are either not-shared publicly or shared in non-standard formats, which makes it difficult to reuse and integrate models.
In turn this causes disbelief of many stakeholders in credibility of models. This subgroup intends to explore these issues and recommend potential remedies. Those issues include:
- Repeatability and Reproducibility of models in publications and repositories
- Credibility of models
- Standardization efforts & specifications for:
     1) Models and modeling technologies.
     2) Modeling data
- Model annotation

Known specific issues:
Models are written in different languages
Models require adaptation towards integration
Models are not licensed in an easy way that allows reuse
Models are hard to locate many times as those are published in different sources despite many repositories available



-----------------------------------------------------------------------
Rahuman Sheriff PhD
Project Leader
European Bioinformatics Institute 
European Molecular Biology Laboratory (EMBL-EBI)
Welcome Trust Genome Campus
Hinxton Cambridge CB10 1SD, UK
Email: sheriff at ebi.ac.uk
-----------------------------------------------------------------------



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.simtk.org/pipermail/vp-reproduce-subgroup/attachments/20210120/ee91045a/attachment.html>