flDPnn: Accurate intrinsic disorder prediction with putative propensities of disorder functions

Abstract

Identification of intrinsic disorder in proteins relies in large part on computational predictors, which demands that their accuracy should be high. Since intrinsic disorder carries out a broad range of cellular functions, it is desirable to couple the disorder and disorder function predictions. We report a computational tool, flDPnn, that provides accurate, fast and comprehensive disorder and disorder function predictions from protein sequences. The recent Critical Assessment of protein Intrinsic Disorder prediction (CAID) experiment and results on other test datasets demonstrate that flDPnn offers accurate predictions of disorder, fully disordered proteins and four common disorder functions. These predictions are substantially better than the results of the existing disorder predictors and methods that predict functions of disorder. Ablation tests reveal that the high predictive performance stems from innovative ways used in flDPnn to derive sequence profiles and encode inputs. flDPnn’s webserver is available at http://biomine.cs.vcu.edu/servers/flDPnn/

DisProt in 2022: improved quality and accessibility of protein intrinsic disorder annotation

“…Recently, it has received a boost with the establishment of the Critical Assessment of Protein Intrinsic Disorder prediction (CAID) experiment as a community-wide blind test to compare state-of-the-art approaches to predict disorder ( 4 ). As new disorder prediction methods keep emerging ( 5–7 ), CAID takes on monitoring the field in real time, aiming to establish dependable standards. This ambition has a special caveat, as predicting and identifying regions in IDPs/IDRs that engage in functional interactions remains a significant challenge ( 6 , 8 ).…”

BP180/Collagen XVII: A Molecular View

“…This disordered nature and potency to fold may demonstrate that interacting macromolecules such as anionic membrane lipids and acidic interacting proteins may be required for the folding of BP180, suggesting that the ICD might be relatively unstable even as a part of the HD complex. This hypothesis is supported by a new model incorporating a recently published flDPnn algorithm [37] which predicts protein and/or nucleic acid interactions with disordered regions of the ICD of BP180 (Figure 3c). It seems that the under-investigated field of intrinsic disorder in keratinocyte proteomes is gathering pace [38].…”

Real-Time Structure Search and Structure Classification for AlphaFold Protein Models

“…We used two methods, flDPnn 18 and SPOT-Disorder-Single 17 . flDPnn uses profile information computed by three other methods, which is processed by a deep learning architecture to output residue-wise disorder prediction.…”

“…Particularly, we analyzed correlation between the low-confidence regions (pLDDT 0.5 and 0.7) from AlphaFold2 models and disorder predictions. We used two disorder prediction methods, SPOT-Disorder-Single 17 and flDPnn 18 . According to the two methods, about 14% to 18% of residues are disordered (Fig.…”