Avoiding Regions Symptomatic of Conformational and Functional Flexibility to Identify Antiviral Targets in Current and Future Coronaviruses

Rahaman, Jordon; Siltberg-Liberles, Jessica

Intrinsic disorder propensity was inferred using two different predictors: IUPred (default settings; "long" option) (Dosztá nyi et al. 2005a (Dosztá nyi et al. , 2005b and DISOPRED2 (Ward et al. 2004 ) for all proteins. For IUPred, the site-specific continuous disorder propensities for each protein were mapped onto their corresponding position in the multiple sequence alignment as raw disorder propensities and as binary states, order or disorder, using two cutoffs of 0.4 and 0.5. Disorder propensities below the cutoff were assigned order and disorder propensities at the cutoff or above were assigned disorder. For the DISOPRED2 predictions that were inferred using the nr database, the continuous disorder propensities for every site in a protein were mapped onto their corresponding position in the multiple sequence alignment as raw disorder propensities and as binary states, order or disorder, using a cutoff of 5. Consequently, for every protein family (a multiple sequence alignment and its corresponding phylogenetic tree), two continuous matrices and three binary matrices resulted: IUPred 0.4, IUPred 0.5, and DISOPRED2. An additional matrix was generated to indicate sites where the binary order and disorder assignments differ between IUPred 0.4 and DISOPRED2.