Incorrect determination of interacting pairs

[a-r-j]

Hi @anton-bushuiev, nice work and congrats on the ICLR acceptance.

I noticed there are some instances of incorrect interfaces in the raw set e.g. 7lgq_D_H.pdb which only contains a single atom for chain H. Is this intended?

[anton-bushuiev]

Hi @a-r-j, thank you for writing!

It is intended that 7lgq_D_H.pdb has only one atom for chain H because in the original PDB file there is only one ATOM record for chain H and the raw set contains all the interactions where at least one contact between different chains is present.

Nevertheless, what is not expected is that the buried surface area in the 7lgq_D_H.pdb file is more than 500A^2 which is a criterion for filtering out "small" interactions like this. I will investigate this further after ICLR. Thank you for catching this!

Are there any other instances that you find incorrect? I can take a closer look.