Skip to main content

Table 4 Prediction rules between classes in SCOP and PROSITE obtained by cross-training

From: A method for probabilistic mapping between protein structure and function taxonomies through cross training

Rules: Structural-properties (SCOP class) ⇒ Functional-properties (PROSITE class)

All α Proteins.Globin like → Globinsfamily profile 0.323

All α Proteins.Globin like → Legume lectins signatures 0.079

All β Proteins.Cupredoxin like → Type-1 copper blue proteins signature 0.200

All β proteins.Galactose-binding domain-like →C-type lectin domain signature & profile 0.182

All β Proteins.Nucleoplasmin-like VP viral coat & capsid→ DEAD & DEAH box families ATP-dependent helicases signatures 0.14

α &β proteins.NADP- binding Rossmann-fold domains → C-type lectin domain signature & profile 0.103

All β proteins.Double-stranded β-helix → EF-hand calcium-binding domain 0.086

All α Proteins.EF Hand like → EF-hand calcium-binding domain 0.090

All β Proteins.Immunoglobulin-like β sandwich → Immunoglobulins & MHC proteins 0.034

All β Proteins.Immunoglobulin-like β sandwich → Ig-like domain profile 0.064

α &β Proteins.α β-Hydrolases → EF-hand calcium-binding domain 0.086

α &β proteins.Ferredoxin-like → 4Fe-4S ferredoxins iron sulfur binding region signature 0.024

α &β proteins. P-loop containing nucleoside triphosphatehydrolases → Heavy-metal-associated domain signature & profile 0.037

α &β proteins.Thioredoxin fold → Pyridine nucleotide-disulphide oxidoreductases class-I active site 0.097

All α proteins.Four-helical up-&-down bundle → Cytochrome c family heme-binding site signature 0.094

Membrane & cell surface proteins and peptides → C-type lectin domain signature & profile 0.092

All β proteins.Concanavalin A-like lectins glucanases → C-type lectin domain signature & profile 0.084

All α proteins.DNA RNA-binding 3-helical bundle → TNFR NGFR family cysteine-rich region signature & profile 0.062

Small proteins → Serine proteases trypsin family signatures & profile 0.010

Small proteins → Ig-like domain profile 0.047

Rules: Functional-properties (PROSITE class) ⇒ Structural-properties (SCOP class)

Globins family profile → All α proteins.Globin-like 0.369

Serine proteases trypsin family signatures & profile → Small proteins 0.722

Protein kinases signatures & profile → Small proteins 0.662

Legume lectins signatures → All α proteins.Globin-like 0.060

Cytochrome c family heme-binding site signature → All β proteins.Barrel-sandwich hybrid 0.029

  1. Selected prediction rules found as a by-product of cross-training SCOP and PROSITE. Top panel shows rules probabilistically predicting the PROSITE domain/signature given the SCOP superfamily, while the bottom panel shows rules predicting SCOP superfamily given the PROSITE domains. DOT ('.') links parent class to child class in SCOP hierarchy. The bold value following each rule is the probabilistic weighted score found using cross training.