Length-dependent prediction of protein intrinsic disorder

By Kang Peng, Predrag Radivojac, Slobodan Vucetic, A. Keith Dunker, Zoran Obradovic. 2006

915 citations recorded by the source corpus.

BackgroundDue to the functional importance of intrinsically disordered proteins or protein regions, prediction of intrinsic protein disorder from amino acid sequence has become an area of active research as witnessed in the 6th experiment on Critical Assessment of Techniques for Protein Structure Prediction (CASP6). Since the initial work by Romero et al. (Identifying disordered regions in proteins from amino acid sequences, IEEE Int. Conf. Neural Netw., 1997), our group has developed several predictors optimized for long disordered regions (>30 residues) with prediction accuracy exceeding 85%. However, these predictors are less successful on short disordered regions (≤30 residues). A probable cause is a length-dependent amino acid compositions and sequence properties of disordered regions.ResultsWe proposed two new predictor models, VSL2-M1 and VSL2-M2, to address this length-dependency problem in prediction of intrinsic protein disorder. These two predictors are similar to the original VSL1 predictor used in the CASP6 experiment. In both models, two specialized predictors were first built and optimized for short (≤30 residues) and long disordered regions (>30 residues), respectively. A meta predictor was then trained to integrate the specialized predictors into the final predictor model. As the 10-fold cross-validation results showed, the VSL2 predictors achieved well-balanced prediction accuracies of 81% on both short and long disordered regions. Comparisons over the VSL2 training dataset via 10-fold cross-validation and a blind-test set of unrelated recent PDB chains indicated that VSL2 predictors were significantly more accurate than several existing predictors of intrinsic protein disorder.ConclusionThe VSL2 predictors are applicable to disordered regions of any length and can accurately identify the short disordered regions that are often misclassified by our previous disorder predictors. The success of the VSL2 predictors further confirmed the previously observed differences in amino acid compositions and sequence properties between short and long disordered regions, and justified our approaches for modelling short and long disordered regions separately. The VSL2 predictors are freely accessible for non-commercial use at http://www.ist.temple.edu/disprot/predictorVSL2.php

Subfields: Materials Chemistry, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

In Molecular Biology

Citations in the corpus

Programmable protein delivery with a bacterial contractile injection system (2023, positive): “The N-terminal domain (NTD) of Pdp1 is likely disordered as it exhibits a low pLDDT score in AlphaFold (it appears as a long disordered extension in the protein’s predicted structure) as well as a high PONDR VSL2 score 49 .”
SPOT-Disorder2: Improved Protein Intrinsic Disorder Prediction by Ensembled Deep Learning (2019, methodology): “See above-mentioned references [64–66 , 68– 70 ] for further information. of We further employed other independent test datasets to compare our methods with other top performing methods for DisProt228 that are available to us as either a local implementation or online server.”
UniRef: comprehensive and non-redundant UniProt reference clusters (2007, neutral): “…take advantage of the reduced sampling bias in UniRef to develop and improve profile-based models (PSI-BLAST, HMMER) for specific gene and structural motifs (Casbon and Saqi, 2006; Kinjo and Nishikawa, 2006; McGuffin et al., 2006; Peng et al., 2005; Rojas et al., 2005; Wang and Samudrala, 2006).”
Improving protein disorder prediction by deep bidirectional long short‐term memory recurrent neural networks (2016, methodology): “Examples are CSpritz (Walsh et al. , 2011), MD (Schlessinger et al. , 2009), metaPrDOS (Ishida and Kinoshita, 2008), MFDp2 (Mizianty et al. , 2013) and PONDR-FIT (Xue et al. , 2010).”
Bioinformatical approaches to characterize intrinsically disordered/unstructured proteins (2010, methodology): “This architecture ensured that PONDR VSL2 has a more balanced performance on disordered segments of various lengths.”
Comprehensive review of methods for prediction of intrinsic disorder and its molecular functions (2017, methodology): “PONDR-VSL2B 2006 [52, 53] WS ? SP No Yes http://www.”
Natively unstructured regions in proteins identified from contact predictions (2007, methodology): “Another example is a meta-method that uses two different prediction methods, each optimized on unstructured regions of different lengths ( Peng, et al., 2006 ).”
Evaluation of disorder predictions in CASP9 (2011, background): “The issue of the different length of disordered regions has been taken into consideration in several disorder predictions methods.49 –53 To address this issue in the assessment, we additionally evaluated the predictions taking into account only segments longer than a specified length cutoff.”
Comprehensive comparative assessment of in-silico predictors of disordered regions. (2012, methodology): “predictors built utilizing machine learning classifiers, such as DISOPRED [[24]], DisEMBL [[25]], DISOPRED2 [[26]], DISpro [[27]], RONN [[28]], Spritz [[29]], ProfBval [[30],[31]], PONDR family of predictors [[10],[32]-[36]], DisPSSMP [[37]], DisPSSMP2 [[38]], POODLE family of predictors [[39],[40]]”
A tissue-specific atlas of mouse protein phosphorylation and expression. (2010, positive): “Indeed, examination of predicted structural elements within proteins using PsiPred (Jones, 1999) and VSL2 (Peng et al., 2006) revealed that phosphorylated Ser, Thr, and Tyr residues exhibited marked differences in structural classification compared to their unmodified counterparts (Figure S3E).”
Performance of mutation pathogenicity prediction methods on missense variants (2011, methodology): “…transmembrane helix prediction by TMHMM [Krogh et al., 2001], coiled-coil structure prediction by MARCOIL [Delorenzi and Speed, 2002], stability prediction by I-Mutant 2.0 [Capriotti et al., 2005], B-factor prediction [Radivojac et al., 2004], and disorder prediction by DisProt [Peng et al., 2006].”
Protein tandem repeats: the more perfect the less structured (2010, methodology): “…disorder-promoting residues, we further analysed the abundance of predicted intrinsic disorder in these sequences with several computational tools, including the pondr (cid:2) vlxt [34,40] and vsl 2 [41,42] algorithms, as well as predictors such as iupred [43,44], foldindex [45], and topidp [37].”