Some more features

[shrimonmuke0202]

I want the volume, dipole moment, hydrogen bond donor, and hydrogen bond acceptor for the node feature in the protein residue graph. How can I do this using graphein ?

[a-r-j]

Hiya,

Could you clarify what you mean by volume?

Dipole moments aren’t currently implemented, but you can use a tool like this to compute them.

For HBA/HBD, it should be pretty straightforward to define your own function based on this table:

Amino acids	Hydrogen donor atoms	Hydrogen acceptor atoms
Arginine (Arg, R)	NE, NH1 (2), NH2 (2)
Asparagine (Asn, N)	ND2 (2)	OD1 (2)
Aspartic acid (Asp, D)		OD1 (2), OD2 (2)
Glutamine (Gln, Q)	NE2 (2)	OE1 (2)
Glutamic acid (Glu, E)		OE1 (2), OE2 (2)
Histidine (His, H)	ND1, NE2	ND1, NE2
Lysine (Lys, K)	NZ (3)
Serine (Ser, S)	OG	OG (2)
Threonine (Thr, T)	OG1	OG1 (2)
Tryptophan (Trp, W)	NE1
Tyrosine (Tyr, Y)	OH	OH

Eg.

def hbond_type(n, d):
    If d[“residue_type”] == “Y”:
        d[“hbond_donor”] = True
        D[“hbond_acceptrot = True
    ….

[a-r-j]

Graphein provides featurisation with aaindex - make sure you install the extra dependency in .requirements/extras.in. You can see here for the possible features - there are several volumes. There also seem to be some Hbond-related features you can look at.

[a-r-j]

Hey @shrimonmuke0202 I've added a Hbond donor and acceptor feature in #150 - What do you think? Is this what you had in mind?

[shrimonmuke0202]

Sorry for the late reply; this will help me a lot.

[a-r-j]

Great! Hbond acceptors/donors are now in 1.3.0. Did you manage to get the volume features? I can also look into the dipole moments.

[shrimonmuke0202]

No, I did not get the volume features and the dipole moments.