ABSTRACT

The biological processes and molecular functions associated with protein disorder have been comprehensively addressed in several studies (Table 11.1). Ward and colleagues (Ward et al. 2004) applied DISOPRED2 to predict the frequency of proteins with intrinsically disordered regions (IDRs) ≥30 consecutive residues in 24 genomes from the three kingdoms of life and also the involvement of proteins in different GO annotations in yeast. Tompa and colleagues (Tompa, Dosztanyi, and Simon 2006b) combined the results of IUPred and PONDR® VSL1, and compared the proteomes of E. coli and yeast by several criteria, such as the percentage of all disordered residues, the percentage of fully disordered proteins, and proteins with IDRs ≥30 consecutive residues. Dunker and colleagues (Xie et al. 2007) predicted disorder in SwissProt entries and looked for keywords in their functional annotation record that showed statistically significant associations (238 out of 710 Swiss-Prot functional keywords). This study extended previous work, in which disorder in 12

different functional categories, either related to BP or MF, was directly assessed (Iakoucheva et al. 2002). The bioinformatic analyses are complemented by statistically less rigorous studies, in which a census of functional annotations of IDPs (i.e., MF terms) (e.g., protein-protein binding, protein-DNA binding, metal binding, phosphorylation), and occasionally BP terms (regulation of proteolysis) is given (Dunker et al. 2002). Often, exclusive functional attributes of IDPs not yet incorporated into GO (e.g., flexible linker/spacer, entropic spring, protein detergent; see also Chapter 12) are included in these classification schemes.