Alignments grow, secondary structure prediction improves

Przybylski, Dariusz; Rost, Burkhard

doi:10.1002/prot.10029

Cited by 179 publications

(135 citation statements)

References 61 publications

(98 reference statements)

Supporting

Mentioning

127

Contrasting

Unclassified

Order By: Relevance

“…Furthermore, the VL3E predictor which combi ned VL3H and VL3P resulted in an additional accuracy improvement of more than 3.1% compared to VL3. T hese results are in accord with improvements achieved in secondary structure predictions 42 .…”

Section: Discussionsupporting

confidence: 86%

“…Since the available dataset with 152 disordered proteins in DIS152 was fairly small and likely to constrain the achievable accuracy of disorder prediction, the data set was enhanced by including homologues of the disordered sequences. Based on previous results showing that using evolutionary information as part of the prediction process improved the accuracy of secondary structure prediction [40][41][42][43] , we hypothesized that including homologues in the disorder training set would also improve prediction of protein disorder.…”

Section: Pondr Vl3 Homology (Vl3h)mentioning

confidence: 99%

See 1 more Smart Citation

Optimizing Long Intrinsic Disorder Predictors With Protein Evolutionary Information

Peng

Vučetić

Radivojac

et al. 2005

J. Bioinform. Comput. Biol.

468

428

View full text Add to dashboard Cite

Protein existing as an ensemble of structures, called intrinsically disordered, has been shown to be res ponsible for a wide variety of biological functions and to be common in nature. Here we focus on improving sequence-based predictions of long (> 30 amino acid residues) regions lacking specific 3-D structure by means of four new neural -network-based Predictors Of Natural Disordered Regions (PONDRs): VL3, VL3H, VL3P, and VL3E. PONDR VL3 used several features from a previously introduced PONDR VL2, but benefitted from optimized predictor models and a slightly larger (152 versus 145) set of disordered proteins that were cleaned of mislabeling errors found in the smaller set. PONDR VL3H utilized homologues of the disordered proteins in the training stage, while PONDR VL3P used attributes derived from sequence profiles obtained by PSI-BLAST searches. The measure of accuracy was the average between accuracies on disordered and ordered protein regions. By this measure, the 30-fold cross-validation accuracies of VL3, VL3H, and VL3P were, respectively, 83.6±1.4%, 85.3±1.4%, and 85.2±1.5%. By combining VL3H and VL3P, t he resulting PONDR VL3E achieved an accuracy of 86.7±1.4%. This is a significant improvement over our previous PONDRs VLXT (71.6±1.3%) and VL2 (80.9±1.4% ). The new disorder predictors with the corresponding datasets are freely accessible through the web server at www.ist.temple.edu/disprot.

show abstract

Section: Discussionsupporting

confidence: 86%

Section: Pondr Vl3 Homology (Vl3h)mentioning

confidence: 99%

Optimizing Long Intrinsic Disorder Predictors With Protein Evolutionary Information

Peng

Vučetić

Radivojac

et al. 2005

J. Bioinform. Comput. Biol.

468

428

View full text Add to dashboard Cite

show abstract

“…The filtering procedures resulted in a set of 723 non-redundant disordered chains. To leverage evolutionary information, PSI-BLAST (Altschul et al, 1997) is used to generate profiles by aligning all chains against the Non-Redundant (NR) database, as in (Jones, 1999;Przybylski and Rost, 2002;Pollastri et al, 2001b). Finally, these chains were randomly split into ten subsets of approximately equal size for ten-fold cross-validated training and testing.…”

Section: Datamentioning

confidence: 99%

Accurate Prediction of Protein Disordered Regions by Mining Protein Structure Data

Cheng

Sweredoski

Baldi

2005

Data Min Knowl Disc

195

162

View full text Add to dashboard Cite

Abstract. Intrinsically disordered regions in proteins are relatively frequent and important for our understanding of molecular recognition and assembly, and protein structure and function. From an algorithmic standpoint, flagging large disordered regions is also imortant for ab inito protein structure prediction methods. Here we first extract a curated, non-redundant, data set of protein disordered regions from the Protein Data Bank and compute relevant statistics on the length and location of these regions. We then develop an ab initio predictor of disordered regions called DISpro which uses evolutionary information in the form of profiles, predicted secondary structure and relative solvent accessibility, and ensembles of 1D-recursive neural networks. DISpro is trained and cross validated using the curated data set. The experimental results show DISpro improves the prediction accuracy of disordered regions over previous methods. DISpro is a member of the SCRATCH suite of protein data mining tools available through

show abstract

“…1B), manages to confer a physiologically relevant structure on human AE1-Ct. Moreover, the PHD program (29,30) predicts that the secondary structure of pure AE1-Ct is virtually identical to the AE1 component of GST-AE1-Ct.…”

Section: Elisa Interactions Between Slc4-ct Domains and Immobilized mentioning

confidence: 99%

Evidence against a Direct Interaction between Intracellular Carbonic Anhydrase II and Pure C-terminal Domains of SLC4 Bicarbonate Transporters

Piermarini

Kim

Boron

2007

Journal of Biological Chemistry

View full text Add to dashboard Cite

Based on solid-phase binding assays with enzyme-linked immunosorbent assay detection, previous investigators suggested that intracellular carbonic anhydrase II (CA II) interacts at high affinity with the C-terminal (Ct) domains of SLC4 bicarbonate-transport proteins, expressed as glutathione S-transferase (GST) fusion proteins, to form functional HCO 3 ؊ metabolons. Here we re-evaluated this protein-protein interaction using two solid-phase binding assays. We first compared the ability of the Ct domain of three SLC4 transporters, SLC4-A1 (AE1), SLC4-A4 (NBCe1), and SLC4-A8 (NDCBE), to bind immobilized CA II, using enzyme-linked immunosorbent assay detection. We found that when expressed as GST fusion proteins, all three bind to CA II (K d 300 -600 nM) better than does pure GST. However, we detected no binding of pure SLC4-Ct peptides to immobilized CA II. Second, we reversed assay orientation by immobilizing the SLC4-Ct fusion proteins or peptides. We found that more CA II binds to GST than to any of the three GST-SLC4-Ct fusion proteins. Furthermore, we detected no binding of CA II to any of the immobilized pure SLC4-Ct peptides. Finally, we used surface plasmon resonance to detect possible rapid interactions between CA II and the pure peptides. Although we detected acetazolamide binding to immobilized CA II and specific antibodies binding to immobilized SLC4-Ct peptides, we detected no binding of CA II to immobilized SLC4-Ct or vice versa. Thus, although an HCO 3 metabolon may exist, CA II cannot bind directly to pure SLC4-Ct peptides and can bind to GST-SLC4-Ct fusion proteins only when the CA II is immobilized and the fusion protein is soluble, and not vice versa.

show abstract

Alignments grow, secondary structure prediction improves

Cited by 179 publications

References 61 publications

Optimizing Long Intrinsic Disorder Predictors With Protein Evolutionary Information

Optimizing Long Intrinsic Disorder Predictors With Protein Evolutionary Information

Accurate Prediction of Protein Disordered Regions by Mining Protein Structure Data

Evidence against a Direct Interaction between Intracellular Carbonic Anhydrase II and Pure C-terminal Domains of SLC4 Bicarbonate Transporters

Contact Info

Product

Resources

About