MLP_mod.py

# kdArrange

import bpy
from collections import defaultdict


''' 
[x] step 01: first store atoms as {element_name: [co,..], } 
[ ] step 02: generate singular ordered mesh by adding vertices 
             in clumps of element types. (H,H,H,H,H,H,O,O,O,O,O,O..)
[ ] step 03: track start and end index for each element into mapper_obj
[ ] step 04: get surface mesh.
[ ] step 05: for every vertex on surface mesh find closest
             vertex in proxy_ob, and MLP value

'''

# step 01
# attrs = [a for a in dir(o) if a.startswith('bb2')]
attrs = [
    'bb2_objectType',
    'bb2_outputOptions',
    'bb2_pdbID',
    'bb2_pdbPath',
    'bb2_subID'
]

proxy_obj = defaultdict(list)
nstr = '_4GE.001'
objs = bpy.data.objects[nstr].children
for o in objs:
    
    # for a in attrs:
    #    print(a, ':', getattr(o, a))

    # coincides and is a relatively fast lookup
    element_name = o.active_material.name

    co = o.location[:]
    proxy_obj[element_name].append(co)
    
# step 03

mapper_obj = {}
verts = []

for key in sorted(proxy_obj.keys()):
    start = len(verts)
    verts.extend(proxy_obj[key])
    end = len(verts)-1
    mapper_obj[key] = (start, end)

# print(mapper_obj)
# {'C': (0, 97), 'N': (98, 124), 'O': (125, 153)}


# step 02

mesh = bpy.data.meshes.new("mesh_name")
mesh.from_pydata(verts, edges=[], faces=[])
mesh.update()

obj = bpy.data.objects.new("obj_name", mesh)

scene = bpy.context.scene
scene.objects.link(obj)

#     

please forgve my insistence: the element type (C, O, N, H, P...) is not the relevant property.
The properties of each atom depend on its connections. Luckily, there are a relatively small number of possible combinations (at least in first approximation, in biological setting), and these can be attributed to each atom on the basis of the aminoacid to which they belong.
Take for example the case of Nitrogen (N):

• it is present in all aminoacids on the MainChain, and in this configuration (linked to a C and to an O) it takes the value -0.49;
• the terminal N in Arginine, part of a guaniudinium group, have value -0.14;
• in yet another configuration, the N of Lysine has -1.07;

The way to identify each atom is by ithe aminoacid name (ARG, LYS, ALA, GLY ecc in the file pdb), and the atom name (N, NE, NH1 ...)

If you need better explanation, just let me know, I'll try to be clearer: feel free to ask!