The calculated results are inconsistent with those calculated by amber cpptraj

[astomer2]

Dear Developers, I'm counting multiple different peptides binding to a receptor and analyzing the differences in the contacts they make. Firstly, I used cpptraj to calculate a nativecontacts as a reference, because the results calculated using ContactFrequency could not display their residue numbers separately according to peptides and receptors, causing all the structures to be mixed together, even if I set the query and haystack. But this problem is only inconvenient, not fatal. I compared the results calculated by cpptraj and contact_map and found that the results calculated by cpptraj have more contact residue pairs and the contact frequency is much higher. How do I interpret this result?
My code is as follows：

freq = ContactFrequency(traj, query=ligand, haystack=receptor, cutoff=0.7)
contact_lists = []
freq.residue_contacts.most_common_idx()
for i in range(len(freq.residue_contacts.most_common_idx())):
tup = freq.residue_contacts.most_common_idx()[i]
if tup[1] >0.2:
contact_lists.append(list(tup[0])[1])
set(contact_lists)

and my cpptraj code:
nativatecontacts :1-255&!@h= :256-356&!@h= byresidue distance 7 out nc_by_frame.dat mindist maxdist
resout contact_frac_byres.dat \

Relevant results have been placed in the attachment
contact_frac_byres.txt

contact_map_output.txt

[sroet]

Dear @astomer2,

First of all, it has been a while since I looked at this code, and secondly I am not familiar with ccptraj, but I will try to help you find some answers:

I compared the results calculated by cpptraj and contact_map and found that the results calculated by cpptraj have more contact residue pairs and the contact frequency is much higher. How do I interpret this result?

From looking at your result, it seems like ccptraj gives the TotalFrac, as a number between 0 and 100, while contactmap gives you a number between 0 and 1. So looking at that it seems like contactmap both gives higher frequencies and way more pairs (if you only look at the entries where TotalFrac is above 20)

I don't know the explicit reasons why these two list differ, but common causes are (in no particular order):

• different definition of the cutoff (angstrom vs nm, between atoms or centers-of-mass or centers-of-geometry)
• different definitions of when a contact exists between residues (should 1 atom be within the cutoff or all, or half?)
• different handling of possible periodic boundary conditions
• different counting of the residue numbers (are they ordered, do you start at 1, or 0, do you use the loaded labels from the pdb, or generate your own?)

I am also very unsure about the Contacts column of you ccptraj output. It does not seem to mean "number of frames with a contact" as it has no correlation with the TotalFrac column, so my first guess is that there are some mismatches in the definition of "what is a contact between residues"