Recently, I noticed that a user had uploaded a file to the website "DSSR-enabled Innovative Schematics of 3D Nucleic Acid Structures with PyMOL", which DSSR reported as 'no nucleotides found.' Upon visualizing it in PyMOL, the structure appeared to be a single-stranded RNA. Further investigation revealed that while the uploaded file was in PDB format, it did not adhere to the standard naming conventions for nucleotides typically used in RCSB PDB entries. For instance, an A nucleotide extracted from the file had its exocyclic amino group named as N553 instead of the conventional N6 (see below).

Following 3DNA, DSSR uses the atomic coordinates and standard names of base-ring atoms to identify a nucleotide. All known nucleotides share a common six-membered pyrimidine ring, with atoms named consecutively (N1, C2, N3, C4, C5, C6), and purines include three additional atoms (N7, C8, N9). See below for the standard names in Watson-Crick base pairs.

Without proper names for base ring atoms, DSSR is unable to identify nucleotides, resulting in the input structure being reported as 'no nucleotides found.' The same principle applies to amino acids in protein structures, such as specific naming conventions for amino nitrogen (N), carbonyl carbon (C), and alpha carbon (CA).

See also the blog posts "Mapping of modified nucleotides in DSSR" and "Name of base atoms in PDB formats".