CancerPPD 2.0 - Statistics

Structure Type	PDB	AlphaFold	Structure Information NA	Total
Peptide (length ≤ 40)	2	11	4	17
Protein (length > 40)	81	309	11	401

Prediction Limitations for Peptide and Protein Structures

In our efforts to predict the structures of peptides and proteins that were not readily available on Protein Data Bank (PDB), we used various tools such as PepStrMod and I-TASSER. These tools are highly effective for predicting a wide range of structures; however, some specific structures could not be predicted. This was primarily due to the presence of peculiar modifications or the inherent limitations of the tools. While most of the structures were accurately modeled, certain cases involving complex chemical/terminal modifications or unusual configurations fell outside the predictive capabilities of our approach. A total of 1314 structures are available in CancerPPD 2.0, of which 1258 were predicted using PepStrMod or I-TASSER, and 56 were found on PDB. We were not able to predict a total of 298 structures. Below is a breif description of the structures that were not predicted:

Natural Peptides

A few natural peptides structures were not predicted because of the following reasons:

Sequence Not Available: Some entries lacked sufficient sequence information.
Length < 5: Peptides with a length of less than 5 residues were challenging to predict since PepStrMod requires a minimum length of 5 residues and a maximum of 25 residues.

Proteins

The following limitations were encountered for proteins from literature:

Mix Chirality: Although protein structures were predicted using the I-TASSER server, some structures could not be predicted due to mixed chirality, a known limitation of the I-TASSER.

Chemical Modifications

Though PepStrMod was able to predict many chemically modified structures, several peptide structures could not be predicted due to their unique chemical modifications. Below are some examples of the chemical modifications:

2-Aminoisobutyric Acid
Aminohexanoic Acid
Dansylglycine
Selenocysteine
Pyroglutamic Acid
β-3-benzothienyl-1-Ala
cyclohexyl-1-carboxylic acid

Terminal Modifications

Certain terminal modifications also led to prediction challenges. While simple acetylation and amidation were easily predicted using PepStrMod, some unique terminal modifications were not predicted. Below are some examples of the terminal modifications:

Folic Acid (N-terminal)
1-octanoyl (N-terminal)
rhodamine B - GABA (N-terminal)
Biotinylated (N-terminal)
Methoxylation (C-terminal)
Fluoromethyl ketone (C-terminal)
Leucinol (C-terminal) and 1-octanoyl (N-terminal)

tabular summary of the total unpredicted structures

Structure Type	Total Count
Natural Peptides	25
Proteins	3
Modified Peptide Structures	162
Terminal Modifications	108
Total	298