Start from the best ones
Protein structures can be determined with various methods, such as X-ray crystallography, Cryo-EM or NMR. To get the final 3D structures, assumptions have to be made. To get the best possible outcome for in silico methods, it is important to start from the best possible structures. Our partners from the Coronavirus Structural Task Force have listed protein targets and explain how to choose the best ones (https://insidecorona.net/for-users/). In addition, not all binding sites might be visible in the static 3D structures. Our partners from Folding@home are working on the discovery of cryptic binding sites and energy minimizations of already resolved structures, so check frequently with them (https://foldingathome.org/covid19/). Also, current targets, models, simulations, and more are being gathered by MolSSI (https://covid.molssi.org/). Lastly, a protein thesaurus is available from CAS (https://www.cas.org/covid-19-protein-target-thesaurus). Beware that most structures do not include glycosylation, which could be important depending on the target.
update (14/05/2020): Please check vlog 2 with detailed information by the Coronavirus Structural Taskforce (Dr. Andrea Thorn) here.
Chemical compound libraries
Make sure there are > 1 billion compounds!
We encourage everyone to start from the largest possible compound libraries. A good starting point is the Zinc15 library (), update: the Zinc15 library in ready-to-dock format https://virtual-flow.org/virtualflow-version-zinc15-library, and the REAL library in ready-to-dock format https://virtual-flow.org/real-library. In addition, we require the SWEETLEAD library to be added (https://simtk.org/projects/sweetlead). CAS made an antiviral library available especially for COVID19 (https://www.cas.org/covid-19-antiviral-compounds-dataset). Moreover, our partner Merck has identified over millions of compounds that are not in public libraries. Please join our platform and ask the program manager "Hermans" to share the Merck library with you.
CPU / GPU requests
We are providing computational resources through our partners: GENCI (French), PRACE (EU), and others to be announced soon. Core hours, either CPU or GPU, are available. However, a short proposal is needed to estimate the total number of core hours to be allocated, to check the viability of the approach, and to make sure the outcome is reliable.
For very large requests (> 1 million core hours) with at least one researcher based in Europe, please submit to PRACE directly (https://prace-ri.eu/prace-support-to-mitigate-impact-of-covid-19-pandemic/).
For French teams only, contact Stephane Requena directly (http://www.genci.fr/fr/node/1036)
For non-EU teams, first check your local supercomputing network. We have limited capacity through our commercial partners (to be announced)
See also a list of resources (https://erf-aisbl.eu/research-infrastructures-offer-for-research-on-covid-19/).