 |
 |
Dear Developers, incidentally I stumbled upon a discrepancy between refmac and phenix I cannot quite explain. Wondering I am, how programs add riding hydrogen atoms to histidines. </end_yoda> from what I see, I gather that protonation is always assumed (4 H-sticks protruding from the ring, and the 'heavy' atom bond lengths agree in both programs within 0.01A with the protonated delocalized structure small molecule values). The difference is that in phenix all N-H are @ 0.86A while refmac generates H's with 0.86 for NE2-HE2 but 0.84A for ND1-HD1. Relative, this is a much larger discrepancy than between any of the 'heavy' atom distances. Is this N-H bond difference valid - the 2 H atoms are not entirely equivalent even in the protonated resonance structure (the respective protonated NCN distances yet remain almost identical), but the difference in the respective N-Hs seems significant. I also note than older versions of coot flagged the ND1-HD1 distance of 0.86 as an outlier, but at least from 4459 on that squawk disappears. Is it possible that some library discrepancy that coot has already corrected remains in refmac (5.7.0032) as the source of writing different N-H distances? Could the resp. fellows (phenix, refmac) briefly comment on the riding H divining? Thx, BR PS images with numbers available offline if needed.