Table 14 Summary of base features

Propensity score

The propensity score is derived from a scoring function that sums the log-odd ratios of the amino acids in the spatial neighborhood (defined in [9]) around each residue in a given protein.

[9]

Residue accessibility

Using NACCESS to calculate the accessibilities of the whole molecule submitted in a pdb file. NACCESS calculates the atomic accessible surface defined by rolling a probe around a van der Waals surface. The residue accessibilities are categorized into 4 classes: all-polar, nonpolar, total-side, and main-chain.

[31]

Secondary structure

Secondary structure refers to highly regular local sub-structures defined by patterns of hydrogen bonds between the main-chain peptide groups.

[26]

In such cases, the chain of amino acids folds into regular repeating structures, such as α helix, β structure, and coil.

Accessible surface area

Calculated using Gerstein et al.’s calc-surface program to measure the accessible surface area of a sphere, on each point of which the center of a solvent molecule can be placed in contact with this atom without penetrating any other atoms of the molecule.

[38],[39]

Atom volume

Calculated using Gerstein et al.’s calc-volume program. It calculates volumes by applying a geometric construction called Voronoi polyhedra to divide the total volume among the atoms in a protein model.

[37]

B factor

The B factor is also known as the Debye-Waller factor or the temperature factor. It is used to describe the attenuation of x-ray scattering or coherent neutron scattering caused by thermal motion. Two B factors of a protein were considered in this study: the B factor of side chain and the B factor of main chain.

[32],[33]

Solvent excluded surface

Calculated using Sanner et al.’s MSMS program, which builds the solvent excluded surface based on the reduced surface.

[34]

Solvent accessible surface

Calculated using Sanner et al.’s MSMS program, which builds the solvent accessible surface based on the reduced surface.

[34]

PSSM

Using PSI-BLAST to search the non-redundant protein database, and derive the information content from a position specific scoring matrix as the base feature.

[36]

Side chain polarity

The 20 amino acids were divided into four categories: polar, nonpolar, acidic polar, and basic polar.

[40]

Hydropathy index

Kyte and Doolittle devised the hydopathy index by applying a sliding-window strategy that continuously determined the average hydopathy in a window as it advanced through the sequence.

[41]

Antigenic propensity

Kolaskar and Tongaonkar analyzed 156 antigenic determinants (<20 residues per determinant) in 34 different proteins to obtain the antigenic propensities of amino acid residues.

[42],[43]

Flexibility

Karplus and Schulz developed the flexibility scale based on the mobility of the protein segments on 31 proteins with known structures.

[35]

Hydrophilic scale

Parker et al. developed the hydrophilic scale based on the high-performance liquid chromatography (HPLC) peptide retention data.

[26]