About 'uniprot.human.scratch_outputs.w_sps.tab_corrected' file about deepaffinity HOT 2 CLOSED

Thank you for your interest in DeepAffinity. 'protein_sequence_SPS_mapping' Is prepared by ourself and used for the paper. It is for proteins in bindingdb so besides other species it is a subset for human proteome. ‘uniprot.human.scratch_outputs.w_sps.tab_corrected’ is kindly prepared by Dr. Tomas Babak at Queens University in 2021. It included all human proteome and was done after the publication.
We didn’t compare the two files to see whether they are consistent in the intersection. So thank you for the note! About your example, it appears that some secondary structure elements’ predicted overall secondary structure type or solvent exposure were not consistent, possibly due to different versions of SCRATCH for the corresponding prediction. Their source amino acid sequences in this case seem to be the same uniprot sequence though.

from deepaffinity.

Thank you so much!

from deepaffinity.