De novo structure modeling for nucleic acids in cryo-EM maps using deep learning

Here, we store the detection results and modeled structures generated by CryoREAD for its testing dataset, consisting of 68 maps. 
Results from different EM-Maps are saved in sub-directories with the name [EMD-ID].
In each subfolder, the modeled structures are saved in [EMD-ID].pdb format, while the detection results are saved in the "2nd_stage_detection" directory. Specifically, we keep the detection outputs of our 2nd stage network, which include the probabilities of phosphate, sugar, base, protein, and the individual bases A, U/T, C, and G.

If you used CryoREAD, please cite: "Xiao Wang, Genki Terashi & Daisuke Kihara. De novo structure modeling for nucleic acids in cryo-EM maps using deep learning. bioArxiv, 2022."