Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Predictive”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,099 records · Page 61Linked to original sources
PubMed2006

Predicting medical and surgical complications of carotid endarterectomy: comparing the risk indexes.

BACKGROUND: Several generic cardiac risk assessment tools predict perioperative cardiac complications, but their ability to predict a broader range of medical, neurologic, and surgical complications is unknown. METHODS: A multicenter retrospective observational cohort study of 1998 patients undergoing carotid endarterectomy (CEA). Complications within 30 days of surgery were assessed by medical record review, including death or nonfatal stroke and cardiac, noncardiac medical, minor neurologic, and wound complications. Logistic regression and receiver operating characteristic curve analyses assessed the predictive abilities of the Goldman, Detsky, Revised Cardiac Risk, and American Society of Anesthesiologists indexes and of 2 CEA-specific risk models (the Halm and Tu scores). RESULTS: Death or stroke occurred in 3.2% of patients, cardiac complications in 4.0%, noncardiac medical complications in 3.2%, minor neurologic complications in 6.9%, and wound complications in 6.0%. All risk models (except the Tu score) significantly predicted cardiac complications equally well (P<.05). All 6 risk models were equivalent in predicting noncardiac medical complications. Only the Revised Cardiac Risk Index and the 2 CEA-specific risk models (Halm and Tu scores) predicted death or stroke and minor neurologic and wound complications. The Halm score was superior in predicting death or stroke compared with the Tu score and the Revised Cardiac Risk Index (area under the receiver operating characteristic curve, 0.72 vs 0.62 and 0.61, respectively; P<.05). Patients with cardiac, noncardiac medical, minor neurologic, or wound complications had 3- to 16-fold increased odds of death or stroke. CONCLUSION: The Halm score CEA-specific risk model and the generic Revised Cardiac Risk Index predicted a broad range of medical, neurologic, and surgical complications following CEA.

Aged↗
PubMed1985

Prediction of growth and development in intensive care nursery graduates at 12 months of age.

Although many perinatal events have been linked with the outcome status of neonatal intensive care unit (NICU) graduates, few studies have evaluated the cumulative longitudinal prediction of outcome. This study followed up 65 term and 139 premature NICU graduates to 12 months' chronologic age. Variables that were utilized in predicting 12-month growth and the neurologic and developmental outcome were maternal and neonatal medical factors, prior growth measurements, and neonatal behavior, which was measured with the Brazelton Scale for term infants and with a modified version for premature infants. The neurologic status was normal in 141 (71.9%) of 204 infants. The average Bayley Scale Mental Developmental Index was 112.7, and the average Psychomotor Developmental Index was 97.5, which was corrected for early gestation. Predictions of 12-month neurologic and developmental status were weak and had been derived only by variables from the neonatal behavioral examination; endurance predicted neurologic status and the motor cluster predicted cognitive and motor development. The prediction of growth at 12 months was high and was derived from prior growth parameters. Thus, although predictions of neurodevelopmental status at 12 months' chronologic age were low, the variables that aided in predicting were from the neonatal behavioral evaluation. This finding provides support for recommending that the functional status of the NICU graduate as well as the more traditional list of perinatal problems be considered when contemplating the infants' short-term outcome.

Child Development↗
PubMed1991

Value of subject height in predicting lower esophageal sphincter location.

Subject height and lower esophageal sphincter location were determined in 213 children and adults to determine whether the anthropometric variable could be used to accurately predict sphincter location across all age ranges. The upper margin of the lower esophageal sphincter was determined with a nasally placed manometry catheter. Height was highly predictive of lower esophageal sphincter location across all subject groups (r2 = .96) and in the youngest subset of subjects (< or = 2 years of age, r2 = .88). The predictive ability of height remained significant but progressively decreased in the four older subject groups (> 2 and < or = 10 years of age, r2 = .74; > 10 and < or = 20 years of age, r2 = .66; > 20 and < or = 40 years, r2 = .58; and > 40 years, r2 = .49). The regression equation that described subjects 2 years of age or younger (L = 0.22[H] + 4.92, where L is the location in centimeters from the nares and H is the height in centimeters) correctly predicted lower esophageal sphincter location within 1.0 cm in 90% of these subjects. In the older subject groups, predicted lower esophageal sphincter location was in error by greater than 2 cm in 25% to 35% of the subjects, even when age group-specific regression equations were used. Decreased predictive ability related to both increasing age and increasing height. We conclude that lower esophageal sphincter location can be predicted from height in subjects up to 2 years of age. The prediction is sufficiently accurate in this age group to allow placement of pH probes without manometric measurements.

Adolescent↗
PubMed1997

Better 1D predictions by experts with machines.

Accuracy of predicting protein secondary structure and solvent accessibility has been improved significantly by using evolutionary information contained in multiple sequence alignments. For the second Asilomar meeting, predictions were made automatically for all targets using the publicly available prediction service PredictProtein. Additionally, a semiautomatic procedure for generating more informative alignments was used in combination with the PHD prediction methods. Results confirmed the estimates for prediction accuracy. Furthermore, the more informative alignments yielded better predictions. The fairly accurate predictions of 1D structure were successfully used by various groups for the Asilomar meeting as first step toward predicting higher dimensions of protein structure.

Expert Systems↗
PubMed1998

GLASS: a tool to visualize protein structure prediction data in three dimensions and evaluate their consistency.

When a protein sequence does not share any significant sequence similarity with a protein of known structure, homology modeling cannot be applied. However, many novel and interesting methods, such as secondary structure prediction, fold recognition, and prediction of long-range interactions, are being developed and have been shown to be reasonably successful in predicting protein structures from sequence data and evolutionary information. The a priori evaluation of the correctness of a prediction obtained by one of these methods is however often problematic. Consequently, it is important to use all available information provided by as many different methods as possible and all the available experimental data about the protein of interest, since the consistency of the results is indicative of the reliability of the prediction. Hence the need has arisen for suitable tools able to compare results provided by different methods and evaluate their consistency. We have therefore constructed GLASS, a general platform to read, visualize, compare, and evaluate prediction results from many different sources and to project these prediction results into three dimensions. In addition, GLASS allows the comparison of selected parameters calculated for a model with the distribution observed in real protein structures, thus providing an easy way to test new methods for evaluating the likelihood of different structural models. GLASS can be considered as a "workbench" for structural predictions useful to both experimentalists and theoreticians.

Amino Acid Sequence↗
PubMed1998

Accuracy of side-chain prediction upon near-native protein backbones generated by Ab initio folding methods.

The ab initio folding problem can be divided into two sequential tasks of approximately equal computational complexity: the generation of native-like backbone folds and the positioning of side chains upon these backbones. The prediction of side-chain conformation in this context is challenging, because at best only the near-native global fold of the protein is known. To test the effect of displacements in the protein backbones on side-chain prediction for folds generated ab initio, sets of near-native backbones (< or = 4 A C alpha RMS error) for four small proteins were generated by two methods. The steric environment surrounding each residue was probed by placing the side chains in the native conformation on each of these decoys, followed by torsion-space optimization to remove steric clashes on a rigid backbone. We observe that on average 40% of the chi1 angles were displaced by 40 degrees or more, effectively setting the limits in accuracy for side-chain modeling under these conditions. Three different algorithms were subsequently used for prediction of side-chain conformation. The average prediction accuracy for the three methods was remarkably similar: 49% to 51% of the chi1 angles were predicted correctly overall (33% to 36% of the chi1+2 angles). Interestingly, when the inter-side-chain interactions were disregarded, the mean accuracy increased. A consensus approach is described, in which side-chain conformations are defined based on the most frequently predicted chi angles for a given method upon each set of near-native backbones. We find that consensus modeling, which de facto includes backbone flexibility, improves side-chain prediction: chi1 accuracy improved to 51-54% (36-42% of chi1+2). Implications of a consensus method for ab initio protein structure prediction are discussed.

Models, Chemical↗
PubMed2000

How good is prediction of protein structural class by the component-coupled method?

Proteins of known structures are usually classified into four structural classes: all-alpha, all-beta, alpha+beta, and alpha/beta type of proteins. A number of methods to predicting the structural class of a protein based on its amino acid composition have been developed during the past few years. Recently, a component-coupled method was developed for predicting protein structural class according to amino acid composition. This method is based on the least Mahalanobis distance principle, and yields much better predicted results in comparison with the previous methods. However, the success rates reported for structural class prediction by different investigators are contradictory. The highest reported accuracies by this method are near 100%, but the lowest one is only about 60%. The goal of this study is to resolve this paradox and to determine the possible upper limit of prediction rate for structural classes. In this paper, based on the normality assumption and the Bayes decision rule for minimum error, a new method is proposed for predicting the structural class of a protein according to its amino acid composition. The detailed theoretical analysis indicates that if the four protein folding classes are governed by the normal distributions, the present method will yield the optimum predictive result in a statistical sense. A non-redundant data set of 1,189 protein domains is used to evaluate the performance of the new method. Our results demonstrate that 60% correctness is the upper limit for a 4-type class prediction from amino acid composition alone for an unknown query protein. The apparent relatively high accuracy level (more than 90%) attained in the previous studies was due to the preselection of test sets, which may not be adequately representative of all unrelated proteins.

Amino Acid Sequence↗
PubMed1999

Nuclear morphometry predicts disease-free interval for clinically localized adenocarcinoma of the prostate treated with definitive radiation therapy.

Men treated for prostate cancer often have unexpected outcomes despite predictive models based on stage, grade and prostate-specific antigen (PSA). Previous results have indicated that nuclear morphometry can predict patient outcome in urologic malignancies. Application of this analytical method in prostate cancer treated with radiation therapy is limited. We have evaluated the predictive ability of nuclear morphometry in such patients. Histologic sections from 23 men with clinically localized adenocarcinoma of the prostate treated with radiation therapy were studied. Nuclear morphometric parameters were assessed using a previously described and validated system. Univariate and multivariate logistic regression analyses and a Cox proportional hazards model were used to assess the ability of nuclear morphometric parameters to predict recurrence and disease-free interval. Ten patients had no recurrence with median follow-up of 47. 5 months, while 13 had recurrence. Gleason grade was not predictive of treatment outcome. Pre-treatment PSA data, available for only 11 patients, were predictive of treatment outcome. Several nuclear morphometric parameters predicted recurrence, including upper quartile of suboptimal circle fit and upper quartile of feret-diameter ratio. A prognostic factor score incorporating these 2 parameters was derived, which predicted disease-free interval (p = 0.0014). Int. J. Cancer (Pred. Oncol.) 84:594-597, 1999.

Adenocarcinoma↗
PubMed1997

Predicting rodent carcinogenicity from mutagenic potency measured in the Ames Salmonella assay.

Many in vitro tests have been developed to identify chemicals that can damage cellular DNA or cause mutations, and secondarily to identify potential carcinogens. The test receiving by far the most use and attention has been the Salmonella (SAL) mutagenesis test developed by Ames and colleagues [(1973): Proc Natl Acad Sci USA 70:2281-2285; (1975): Mutat Res 31:347-364], because of its initial promise of high qualitative (YES/NO) predictivity for cancer in rodents and, by extension, in humans. In addition to the initial reports of high qualitative predictivity, there was also an early report by Meselson and Russell [in Hiatt HH et al (1977): "Origins of Human Cancer, Book C: Human Risk Assessment," pp 1473-1481] of a quantitative relationship between mutagenic potency measured in SAL and carcinogenic potency measured in rodents, for a small number of chemicals. However, other reports using larger numbers of chemicals have found only very weak correlations. The primary purpose of this study was to determine whether mutagenic potency, as measured in a number of different ways, could be used to improve predictivity of carcinogenicity, either qualitatively or quantitatively. To this end, eight measures of SAL mutagenic potency were used. This study firmly establishes that the predictive relationship between mutagenic potency in SAL and rodent carcinogenicity is, at best, weak. When predicting qualitative carcinogenicity, only qualitative mutagenicity is useful; none of the quantitative measures of potency considered improves the carcinogenicity prediction. In fact, when qualitative mutagenicity is forced out of the model, the quantitative measures are still not predictive of carcinogenicity. When predicting quantitative carcinogenicity, several possible methods were considered for summarizing potency over all experiments; however, in all cases, the relationship between mutagenic potency predictors and quantitative carcinogenicity is very weak.

Animals↗
PubMed2001

Prediction of individual patient outcome in cancer: comparison of artificial neural networks and Kaplan--Meier methods.

BACKGROUND: There is a great need for accurate treatment and outcome prediction in cancer. Two methods for prediction, artificial neural networks and Kaplan--Meier plots, have not, to the authors' knowledge, been compared previously. METHODS: This review compares the advantages and disadvantages of the use of artificial neural networks and Kaplan--Meier curves for treatment and outcome prediction in cancer. RESULTS: Artificial neural networks are useful for prediction of outcome for individual patients with cancer because they are as accurate as the best traditional statistical methods, are able to capture complex phenomena without a priori knowledge, and can be reduced to a simpler model if the phenomena are not complex. Kaplan--Meier plots are of limited accuracy for prediction because they require partitioning of variables, require cutting continuous variables into discrete pieces, and can only handle one or two variables effectively. CONCLUSIONS: Artificial neural networks are an efficient statistical method for outcome prediction in cancer that utilizes all available powerful prognostic factors and maximizes predictive accuracy. Use of Kaplan--Meier plots for predictions is discouraged because of serious technical limitations and low accuracy.

Forecasting↗
PubMed2002

Long-term outcome in rheumatoid arthritis: a simple algorithm of baseline parameters can predict radiographic damage, disability, and disease course at 12-year followup.

OBJECTIVES: To predict the long-term outcome of rheumatoid arthritis (RA) with respect to radiographic damage, disability, and disease course using baseline variables, and to construct decision trees identifying patients on an individual level at the extremes of the outcome spectrum of these 3 dimensions. METHODS: The 12-year outcome of 112 female RA patients from a prospective inception cohort was assessed by measuring the tertiles of radiographic damage (measured by the modified Sharp/van der Heijde method, SHS), disability (measured by the Health Assessment Questionnaire, HAQ), and severe disease course as defined by patients with either the 33% highest cumulative disease activity (area under the curve of all observed disease activity scores) or the highest tertile of radiographic damage. Patients in the lowest (mild) and highest tertile (severe) of each outcome measure were identified. All baseline parameters known to be associated with each outcome (demographic and socioeconomic parameters; disease duration; disease activity measures; laboratory measures including rheumatoid factor, HLA typing, percentage agalactosyl IgG, functional and radiographic measures) were entered into cross-validated stepwise logistic regression models to find the best predictive combination of baseline parameters for each of the outcomes. Using the results of the logistic regression models, simple decision trees were constructed to categorize patients at an individual level in a particular prognostic group. RESULTS: After 12 years, the lowest and highest tertiles were, respectively, 42.3 and 189 for the SHS and 0.37 and 1.25 for the HAQ. Fifty-five patients had a severe disease course. Mild and severe radiographic damage could be predicted with an accuracy of 90% and 85%, respectively. Mild and severe HAQ could be predicted with an accuracy of 90% and 84%, respectively, and severe disease course with an accuracy of 81%. The baseline variables found to be predictive of all 3 outcome measures were very similar and consisted of combinations of the following baseline parameters: swollen joint count (SJC), Ritchie score, rheumatoid factor (RF), the presence of erosions, and the HAQ score. Additional knowledge of the HLA typing hardly improved the accuracy of the prediction. To predict outcome at the individual level, simple decision trees were constructed using the RF, HAQ, SJC, and presence of erosions at baseline. CONCLUSION: The present study shows that prediction of outcome in long-term RA is possible and can be done using widely available baseline parameters.

Adult↗
PubMed2002

Predicting the home location of serial offenders: a preliminary comparison of the accuracy of human judges with a geographic profiling system.

The accuracy with which human judges, before and after 'training', could predict the likely home location of serial offenders was compared with predictions produced by a geographic profiling system known as Dragnet. All predictions were derived from ten spatial displays, one for each of ten different U.S. serial murderers, indicating five crime locations. In all conditions participants were asked to place an 'X' on each spatial display corresponding to where they thought the offender lived. In the control condition, a comparison was made between the accuracy of these predictions for 21 participants on two separate occasions a few minutes apart. In the experimental condition, between their first and second predictions the 21 participants were given two heuristics to follow--distance-decay and circle hypothesis. Results showed that participants with no previous knowledge of geographic profiling were able to use the two heuristics to improve the accuracy of their predictions. The overall accuracy of the second set of predictions for the experimental group was also not significantly different from the accuracy of predictions generated by Dragnet.

Adult↗
PubMed2004

A tool for predicting breast carcinoma mortality in women who do not receive adjuvant therapy.

BACKGROUND: Among the several proposed risk classification schemes for predicting survival in women with breast carcinoma, one of the most commonly used is the Nottingham Prognostic Index (NPI). The goal of the current study was to use a continuous prognostic model (similar to those that have already been demonstrated to possess greater predictive accuracy than risk group-based models in other malignancies) to predict breast carcinoma mortality more accurately compared with the NPI. METHODS: A total of 519 women who had been treated with mastectomy and axillary lymph node dissection at Memorial Sloan-Kettering Cancer Center (New York, NY) between 1976 and 1979 met the following requirements for study inclusion: confirmation of the presence of invasive mammary carcinoma, no receipt of neoadjuvant or adjuvant systemic therapy, no previous history of malignancy, and negative lymph node status as assessed on routine histopathologic examination. Paraffin blocks were available for 368 of the 519 eligible patients. All available axillary lymph node tissue blocks were subjected to enhanced pathologic analysis. The competing-risk method was used to predict disease-specific death, and the accuracy of the novel prognostic model that emerged from this process was evaluated using the concordance index. Jackknife and 10-fold cross-validation predictions yielded by this new model were compared with predictions yielded by the NPI. RESULTS: Of the 348 women for whom complete data were available, 73 died of disease; the 15-year probability of breast carcinoma-related death was 20%. On the basis of these 348 cases, the authors developed a prognostic model that took patient age, disease multifocality, tumor size, tumor grade, lymphovascular invasion, and enhanced lymph node staining into account, and using competing-risks regression analysis, they found that this new model predicted disease-specific death more accurately compared with the NPI. CONCLUSIONS: The authors have developed a model for predicting breast carcinoma-specific death with improved accuracy. This tool should be useful in counseling patients with regard to their specific need for adjuvant therapy.

Adult↗
PubMed2006

Predicting the transactivation activity of p53 missense mutants using a four-body potential score derived from Delaunay tessellations.

We describe a novel statistical scoring method based on a computational geometry approach to predict the functional impact (transactivation activity) of missense mutations in the DNA-binding domain (DBD) of the tumor suppressor TP53, which is the most frequently mutated gene in human cancer. Residual scores (RS) for each residue were calculated to reflect differences in the compositional preferences of four nearest-neighbor residues between mutant and wild-type proteins. The RS were then combined into a residual score profile (RSP) representing the RS values for all 194 residues in the DBD. Mutants were grouped into functional categories based on their transactivation activities experimentally measured in yeast functional assays using p53-response elements from eight different promoters. While these functional categories showed significant differences in average RS, the latter lacked resolution power to predict the transactivation activities of individual mutants. In contrast, using decision tree models, we found that the RSP predicted transactivation with an accuracy varying between 64.2% and 78.5% depending on the promoter. Lastly, we used the best model to predict the functional outcome of all missense mutants in the DBD of p53 and compared the predictions with their frequency of occurrence in human cancers. We found that mutants predicted as functional (F) accounted for approximately 14% of all missense mutants found in cancers, while mutants predicted as nonfunctional (NF) represented approximately 86% of the mutants. These results show that this computational approach provides a fast and reliable method for predicting the functional impact of p53 mutants associated with cancer.

Algorithms↗
PubMed2000

Use of roentgenography and magnetic resonance imaging to predict meniscal geometry determined with a three-dimensional coordinate digitizing system.

To evaluate and improve on the procedures used by a tissue bank in selecting donor menisci for transplantation, this study was designed to fulfill four objectives: (a) define and quantify a set of independent parameters for describing the geometry of the medial and lateral menisci, (b) determine how well the sizing protocol of the tissue bank (i.e., two transverse roentgenographic measurements obtained from the injured knee or six transverse magnetic resonance imaging measurements obtained from the contralateral knee) predicts the four standard transverse parameters of the menisci, (c) determine if including one additional transverse roentgenographic measurement for each compartment improves the ability of roentgenograms to predict transverse meniscal parameters, and (d) determine if five magnetic resonance imaging measurements at three different meniscal cross sections of the contralateral knee predict the 15 standard cross-sectional parameters of the meniscus in the injured knee. A laser-based, noncontacting three-dimensional coordinate digitizing system was used to determine surface coordinates from which menisci were reconstructed in a computer. For each reconstructed meniscus, four parameters in the transverse plane and five cross-sectional parameters in each of three regions (i.e., anterior, middle, and posterior) were defined, yielding a set of 19 standard parameters to describe the geometry. Through a correlation analysis, these standard parameters were shown to be largely unrelated to one another, thus confirming that the parameters form an independent set describing the three-dimensional geometry of the menisci. The two roentgenographic measurements were poor predictors of transverse standard meniscal parameters, predicting only one of four standard parameters for the medial meniscus and none of four standard parameters for the lateral meniscus with coefficients of determination greater than or equal to 0.5. Including one additional roentgenographic measurement to the tissue bank protocol increased the number of standard transverse parameters predicted to three of four for the medial meniscus and two of four for the lateral meniscus. Magnetic resonance imaging was better than roentgenography for predicting the three-dimensional meniscal geometry. The transverse measurements from magnetic resonance imaging predicted three of four standard transverse parameters for the medial meniscus and all four for the lateral meniscus. With the addition of the cross-sectional measurements by magnetic resonance imaging, seven of 15 standard cross-sectional parameters were predicted for both the medial and lateral menisci. Assuming that a successful clinical outcome depends on how well an allograft matches the size and shape of the original meniscus, magnetic resonance imaging rather than roentgenography should be used for allograft size-matching by tissue banks.

Adult↗
PubMed2000

Predicting motor recovery of the upper limb after stroke rehabilitation: value of a clinical examination.

BACKGROUND AND PURPOSE: Only a few studies have been conducted to predict motor recovery of the arm after stroke. The aims of this study were to identify which clinical variables, assessed at different points in time, were predictive of motor recovery, and to construct useful regression equations. METHOD: One hundred consecutive stroke patients who had an obvious motor deficit of the upper limb were evaluated on entry to the study (two to five weeks post-stroke) and at two, six and 12 months after stroke. The Brunnström-Fugl-Meyer test was used as the outcome measure. Predictors included demographic data, overall disability, clinical neurological features, neuropsychological factors and secondary shoulder complications. RESULTS: In multiple regression analyses, motor performance was invariably retained as the predictive factor with the highest R-square. Other significant predictive variables were overall disability, muscle tone, proprioception and hemi-inattention. Between 53% and 89% of the total amount of variance was accounted for in all selected models. The accuracy of prediction from clinical measurement in the acute phase diminished as the time span of measurement of outcome increased. Similarly, assessment of the variables at two and six months, rather than in the acute stage, resulted in a considerable improvement in the percentage variance explained at 12 months. The highest accuracy was obtained when predictions were made step-by-step in time. CONCLUSIONS: It is possible to predict motor recovery of the upper limb accurately through the use of a few clinical measures. Predictive equations are proposed, the use of which are practicable in both clinical practice and research.

Adult↗
PubMed1998

Fold prediction by a hierarchy of sequence, threading, and modeling methods.

Several fold recognition algorithms are compared to each other in terms of prediction accuracy and significance. It is shown that on standard benchmarks, hybrid methods, which combine scoring based on sequence-sequence and sequence-structure matching, surpass both sequence and threading methods in the number of accurate predictions. However, the sequence similarity contributes most to the prediction accuracy. This strongly argues that most examples of apparently nonhomologous proteins with similar folds are actually related by evolution. While disappointing from the perspective of the fundamental understanding of protein folding, this adds a new significance to fold recognition methods as a possible first step in function prediction. Despite hybrid methods being more accurate at fold prediction than either the sequence or threading methods, each of the methods is correct in some cases where others have failed. This partly reflects a different perspective on sequence/structure relationship embedded in various methods. To combine predictions from different methods, estimates of significance of predictions are made for all methods. With the help of such estimates, it is possible to develop a "jury" method, which has accuracy higher than any of the single methods. Finally, building full three-dimensional models for all top predictions helps to eliminate possible false positives where alignments, which are optimal in the one-dimensional sequences, lead to unsolvable sterical conflicts for the full three-dimensional models.

Algorithms↗
PubMed2002

Use of structure comparison methods for the refinement of protein structure predictions. I. Identifying the structural family of a protein from low-resolution models.

Predicting the three-dimensional structure of proteins is still one of the most challenging problems in molecular biology. Despite its difficulty, several investigators have started to produce consistently low-resolution predictions for small proteins. However, in most of these cases, the prediction accuracy is still too low to make them useful. In the present article, we address the problem of obtaining better-quality predictions, starting from low-resolution models. To this end, we have devised a new procedure that uses these models, together with structure comparison methods, to identify the structural family of the target protein. This would allow, in a second step not described in the present work, to refine the predictions using conserved features of the identified family. In our approach, the structure database is investigated using predictions, at different accuracy levels, for a given protein. As query structures, we used both low-resolution versions of the native structures, as well as different sets of low accuracy predictions. In general, we found that for predictions with a resolution of > or =5-7 A, structure comparison methods were able to identify the fold of a protein in the top positions.

Databases, Protein↗