Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “predictability”

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 1,765 records · Page 98Linked to original sources
PubMed1989

Prediction of post-cardiopulmonary bypass cardiac output.

The ability to predict cardiac output (CO) before termination of cardiopulmonary bypass (CPB) allows identification of potential complications once the patient is off bypass. We have previously demonstrated that CO early after CPB can be reliably predicted by a plot of venous oxygen saturation at various flow rates on CPB, based on in-line monitoring of venous oxygen saturation. In this study, we evaluated a simplified technique for predicting CO with a series of 50 patients on CPB. When CPB weaning began, patients were normothermic, anesthetized, and paralyzed. Venous oxygen saturation and arterial blood flow were recorded. At low pump flow just before termination of CPB, the final venous oxygen saturation was recorded. Assuming a proportional relationship between venous oxygen saturation and arterial blood flow, CO early after CPB was predicted. The simplified CO prediction was compared with the thermodilution CO immediately after CPB. The simplified technique reliably predicted CO early after CPB compared with the thermodilution technique. The simplicity and reliability increase the clinical value of the CO prediction.

Adult↗
PubMed1977

Evaluation of secondary structure predictions in proteins.

Data of 33 proteins are used to compare four methods which predict secondary structure from the amino acid sequence. The prediction of alpha-helices according to the histogram method of Argos et al. (Argos, P., Schwarz, J. and Schwarz, J. (1976) Biochim. Biophys. Acta 439, 261-273) is on the average as reliable as the statistical method of Nagano (Nagano, K. (1973) J. Mol. Biol. 75, 401-420). Both methods predict helices more accurately than the stereo-chemical method of Lim (Lim, V.I. (1974) J. Mol. Biol. 88, 873-894). The method of Nagano yields the best prediction of beta-structure, while the beta-structure predictions of Lim and Argos et al. are not significantly different. The results of the alpha-helix and beta-structure predictions according to the statistical mechanical method of Tanaka and Scheraga (Tanaka, S. and Scheraga, H.A. (1976) Macromolecules 9, 168-182) are inferior to those obtained by the other three methods. For the prediction of turns, no significant difference between the methods of Nagano (Nagano, K. (1974) J. Mol. Biol. 84, 337-372) and Argos et al. was found.

Amino Acid Sequence↗
PubMed1989

Predicting and remembering recurrent pain.

Two groups of Ss suffering from recurrent episodes of pain (headache or menstrual) were studied over a period of weeks in order to determine if they tend to over-predict aversive events (pain in this instance). The results are consistent with earlier findings on fear/pain. Contrary to earlier findings, however, after an under-predicted pain the menstrual Ss decreased their following prediction--suggesting the operation of a superordinate predictive pattern in this group. The Ss in both groups successfully predicted the qualities of their pain episodes, but tended to recall the episodes as being more painful than they had reported at the time of the pain episode. The explanation for the erroneous magnifications of predictions and reports of pain is not obvious. Over-predicting a potential pain, and remembering an episode as having been more painful than it was, may serve a protective function, but if so, these are not the most economical ways of achieving protection. The Ss who used medication expected to obtain moderate relief and they were not disappointed.

Adult↗
PubMed1995

Method of test administration as a factor in test validity: the use of a personality questionnaire in the prediction of cancer and coronary heart disease.

This study examines the predictive accuracy of four different methods of administration of a questionnaire designed to predict cancer and coronary heart disease (CHD) in healthy probands. The method of administration uses the establishment of trust and the explanation of questions as variables in all four possible combinations, i.e. trust and explanation, trust only, explanation only, and neither, the prediction being that the combination of trust and explanation would produce the most accurate prediction, the treatment using neither the worst prediction, with methods using either trust alone or explanation alone intermediate. The criterion was the successful prediction of cancer and CHD. A total population of 3563 men and women was used, and followed up over 15 years, death certificates being used to establish cause of death. As predicted, the combination of trust and explanation did best, use of neither worst. Explanation seemed more important than trust, and the combination seemed to have a synergistic effect. It is apparent that method of administration had an important effect on the outcome of the experiment.

Adult↗
PubMed1988

Time course of changes in serotonin and noradrenaline in rat brain after predictable or unpredictable shock.

The effects of predictable and unpredictable shock on concentrations of serotonin (5-hydroxytryptamine, 5-HT), 5-hydroxyindoleacetic acid (5-HIAA), tryptophan (TP) and noradrenaline (NA) have been studied in 7 regions of rat brain. Two separate experiments have been carried out determining these substances both at 30 min and 2 h after the stress session. Unpredictable shock depleted NA levels in all brain regions except the striatum. However, at 2 h poststress NA in these regions increased significantly in comparison with both controls and predictably shocked rats. Predictable shock also decreased NA in locus coeruleus, brainstem and hypothalamus, which was not observed 2 h later. Both predictable and unpredictable shock decreased 5-HT in brainstem and hypothalamus. At 2 h poststress, 5-HT levels in these regions were still decreased in predictably shocked rats, but had attained control values in unpredictably shocked rats. 5-HT metabolism expressed as the 5-HIAA/5-HT ratio, was significantly increased 30 min after predictable shock in all regions except the locus coeruleus and hippocampus. Unpredictable shock produced a much more marked increase in 5-HIAA/5-HT ratio. At 2 h poststress 5-HT metabolism returned to control values in most of the brain regions of predictably shocked animals, but it remained high after unpredictable shock. The activation of serotonergic metabolism following each type of shock is different according to the nucleus in which the 5-HT nerve endings originate. Only slight increases in tryptophan were observed after both types of shock. Our results suggest that unpredictable shock is perceived as a more anxiogenic situation and that under this condition both 5-HT and NA levels are more effectively normalized with time.

Animals↗
PubMed1996

Prediction of potential protein-protein interaction sites from amino acid sequence. Identification of a fibrin polymerization site.

Identification of a protein-protein interaction site is an important step that has significant potential to clarify structure-function relationships of proteins and drug design. We propose here a unique predictive method to identify protein-protein interaction sites based on the observation that proline is the most common residue found in the flanking segments of interaction sites [Kini, R.M. and Evans, H.J. (1995) Biochem. Biophys. Res. Commun. 212, 1115-1124]. Accordingly, the interaction sites of proteins might be predicted directly from the amino acid sequence based on the presence of proline brackets. Using this strategy, we have predicted a polymerization site in the epitope of the Aalpha-chain of fibrinogen recognized by a monoclonal antibody, 9E9 which inhibits fibrin polymerization [Cierniewski, C.S. and Budzynski, A.Z. (1992) Biochemistry 31, 4248-4253]. The synthetic peptide comprising this predicted site inhibited the coagulation of human blood and allosterically interfered in fibrin polymerization. This is the first known allosteric polymerization site of fibrinogen. Thus the results validate the predicted site and the method for prediction. This unique predictive method should help in identifying the interaction sites of many proteins.

Allosteric Regulation↗
PubMed1992

Prediction of the loading along the leg during snow skiing.

The complete force and moment of each cross section of the leg between the ski boot top and the knee during normal skiing were predicted from measurements of the force and moment under the toe and heel of the boot and the flexion of the ankle. The force and moment components predicted at the base of the boot were significantly different from those predicted at sites of potential injury at the boot top and the knee. The maximum torsional and maximum varus-valgus moments predicted at the knee over all subjects tested were 70 Nm and 149 Nm, which are within the estimated range of the ultimate strength of the knee without support from contracted muscles crossing the knee. Regression analyses were used to find the force components at the base of the boot that best predict the bending and torsional moments at the boot top and knee. The torsional moments at the boot top and knee are best predicted by the medial-lateral force at the toe. The varus-valgus moment at the boot top and knee are best predicted by the resultant medial-lateral force component at the base of the boot. The set of best predictors of the anterior-posterior bending moments at the boot top and knee includes the vertical force at the toe, the vertical force at the heel and the component of the total vertical force directed perpendicular to the leg.

Adult↗
PubMed1992

Predictive assay for rodent carcinogenicity using in vivo biochemical parameters: operational characteristics and complementarity.

111 chemicals of known rodent carcinogenicity (49 carcinogens, 62 noncarcinogens), including many promoters of carcinogenesis, nongenotoxic carcinogens, hepatocarcinogens, and halogenated hydrocarbons, were selected for study. The chemicals were administered by gavage in two dose levels to female Sprague-Dawley rats. The effects of these 111 chemicals on 4 biochemical assays (hepatic DNA damage by alkaline elution (DD), hepatic ornithine decarboxylase activity (ODC), serum alanine aminotransferase activity (ALT), and hepatic cytochrome P-450 content (P450)) were determined. Composite parameters are defined as follows: CP = [ODC and P450), CT = [ALT and ODC), and TS = [DD or CP or CT]. The operational characteristics of TS for predicting rodent cancer were sensitivity 55%, specificity 87%, positive predictivity 77%, negative predictivity 71%, and concordance 73%. For these chemicals, the 73% concordance of this study was superior to the concordance obtained from published data from other laboratories on the Ames test (53%), structural alerts (SA) (46%), chromosome aberrations in Chinese hamster ovary cells (ABS) (48%), cell mutation in mouse lymphoma 15178Y cells (MOLY) (52%), and sister-chromatid exchange in Chinese hamster ovary cells (SCE) (60%). The 4 in vivo biochemical assays were complementary to each other. The composite parameter TS also shows complementarity to all 5 other predictors of rodent cancer examined in this paper. For example, the Ames test alone has a concordance of only 53%. In combination with TS, the concordance is increased to 62% (Ames or TS) or to 63% (Ames and TS). For the 67 chemicals with data available for SA, the concordance for predicting rodent carcinogenicity was 47% (for SA alone), 54% (for SA or TS), and 66% (for SA and TS). These biochemical assays will be useful: (1) to predict rodent carcinogenicity per se, (2) to 'confirm' the results of short-term mutagenicity tests by the high specificity mode of the biochemical assays (the specificity and positive predictivity are both 100%), and (3) to be a component of future complementary batteries of tests for predicting rodent carcinogenicity.

Alanine Transaminase↗
PubMed1995

Predicting chemical carcinogenesis in rodents: the state of the art in light of a comparative exercise.

Within a recent comparative exercise, different approaches to the prediction of rodent carcinogenicity were challenged on a common set of chemicals bioassayed by the U.S. National Toxicology Program. The approaches were of very different natures. Some prediction systems looked for relationships between carcinogenicity and other, more quickly detectable biological events (activity-activity relationships, AAR). Some approaches tended to find structure-activity relationships (SAR). To give an objective evaluation of the results of the exercise, we have analyzed the rodent results and the predictions with the multivariate data analysis methods. The calculated performances varied according to the adopted carcinogenicity classification of the chemicals. When the four rodent results were summarized into a final + or - call, the Tennant approach (AAR method) showed the best performance (about 75% accuracy), whereas the best SAR systems had 60-65% accuracy. A common limitation of almost all the systems was the lack of specificity (too many false positives). Based on these results, better concordance was obtained when the input information was the very costly (and closer to the final endpoint) biological data, rather than the inexpensive (and farther from the endpoint) knowledge of the chemical structure. However, when the rodent results were summarized into a carcinogenicity classification that maintained, to some extent, the gradation intrinsic to the original experimental data, the performance of the AAR systems declined, and the SAR approaches showed a better performance. The difficulty in evaluating the various approaches was further complicated because of a fundamental difference in the approaches themselves: some approaches were 'pure' prediction methods (i.e. their predictions were rigorously based on information not inclusive of carcinogenicity); other approaches (e.g. Tennant, Weisburger) used 'mixed' information, inclusive of known carcinogenicity results from experiments performed before the NTP bioassays. As far as the SAR systems are concerned, their sets of predictions showed a fundamental similarity. This happened in spite of the extremely different procedures adopted to treat the chemical formula (initial information): very simple calculations (Benigni), intuition of the experts (Weisburger and Lijinsky), sophisticated computer programs (TOPKAT and CASE). The results of the Bakale Ke method, based on the experimental measurement of the chemical electrophilicity, and of the Salmonella typhimurium mutagenicity assay were similar to the patterns of predictions of the SAR methods.

Animals↗
PubMed1981

Hearing loss prediction by the acoustic reflex in a young population: comparison of seven methods.

Seven methods of hearing loss prediction by the acoustic reflex were compared in 869 subjects, aged 1 to 20 years. The seven methods were two versions of Sensitivity Prediction by Acoustic Reflex (SPAR), four regression equations, and a bivariate plot technique. While accuracy rate for each of the seven methods generally exceeded 64%, hearing loss was most accurately identified by the bivariate plot system. The proportions of predictive errors among methods, however, ranged from 22 to 31%; moderate errors predominated while severe predictive errors were rare. Accuracy of hearing loss prediction was not significantly affected by chronologic age. However, all methods produced a substantial (31-67%) proportion of predictive errors in patients with minor, clinically insignificant, tympanometric abnormalities. In children with normal impedance findings, the methods permit reasonably accurate identification of hearing loss, categorization of degree of loss, or prediction of hearing loss in dB.

Acoustic Impedance Tests↗
PubMed2026

Risk Factors and Predictive Model for Postoperative High Myopia in Children Undergoing Congenital Cataract Surgery With Intraocular Lens Implantation.

PURPOSE: To identify risk factors associated with the development of high myopia following congenital cataract surgery and to establish a robust predictive model. DESIGN: Retrospective clinical cohort study. SUBJECTS: This retrospective study included 106 pediatric patients who underwent congenital cataract surgery with primary IOL implantation (mean follow-up 8.19 years). The model was externally validated in an independent cohort of 72 patients with a mean follow-up of 7.83 years. METHODS: Preoperative and postoperative ocular biometric parameters were collected. Risk factors for postoperative high myopia were analyzed using Cox proportional hazards regression, which served as the basis for model construction. The predictive performance of the model was rigorously evaluated for discrimination and calibration. Discriminative ability was quantified using Harrell's C-index and the area under the receiver operating characteristic curve (AUC). Model calibration was assessed via calibration plots by comparing predicted probabilities with actual observed outcomes. Internal validation was performed using a bootstrapping method (500 iterations) to ensure model stability and adjust for potential overfitting. RESULTS: An initial postoperative refraction of <+0.75D, and a higher IOL Power to Axial length Ratio (IOL/AL ratio) were identified as significant risk factors for the development of postoperative high myopia. Shorter preoperative axial length was associated with a greater magnitude of postoperative myopic shift. The predictive model demonstrated robust performance, achieving a C-index of 0.711 (internal validation C-index: 0.713). The area under the receiver operating characteristic curve (AUC) values for predicting high myopia at 5 and 10 years were 0.858 and 0.745, respectively. Furthermore, calibration curves demonstrated excellent agreement between the predicted and observed outcomes throughout the follow-up period. In external validation, the model achieved a C-index of 0.825, 5-year AUC of 0.833, and 10-year AUC of 0.713. CONCLUSIONS: Our analysis established that initial postoperative refraction <+0.75D, and an elevated IOL/AL ratio are key determinants of high myopia risk following surgery. Shorter preoperative axial length was associated with a greater magnitude of postoperative myopic shift. This predictive framework provides clinicians with a practical tool to optimize preoperative IOL selection and identify high-risk infants who require vigilant myopia prevention and balanced amblyopia management.

Humans↗
PubMed2005

Comparison of 2-dimensional and 3-dimensional power-Doppler imaging in complex adnexal masses for the prediction of ovarian cancer.

OBJECTIVE: The purpose of this study was to compare 2-dimensional and 3-dimensional power-Doppler imaging diagnostic performance for the prediction of ovarian cancer in complex adnexal masses. STUDY DESIGN: Sixty-nine complex adnexal masses in 60 women (mean age, 48.4 years [range, 17-82 years]) were evaluated by 2-dimensional and 3-dimensional power-Doppler imaging for differentiating benign from malignant masses. Complex adnexal mass was defined in the presence of at least 1 of the following features: solid areas, thick papillary projections, thick septa, or purely solid echogenicity. One examiner performed 2-dimensional power-Doppler imaging, and a second examiner performed 3-dimensional power-Doppler imaging. All masses were removed surgically, and definitive diagnosis was obtained. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated for both techniques. RESULTS: Forty-five tumors (65.2%) were proved to be malignant, and 24 tumors (34.8%) were proved to be benign. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for 2-dimensional power-Doppler imaging were 97.8%, 87.5%, 93.6%, 95.5%, and 94.2%, respectively. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for 3-dimensional power-Doppler imaging were 97.8%, 79.2%, 89.9%, 95%, and 91.3 % respectively. There were no statistical differences in sensitivity and specificity (McNemar test: P = .250) CONCLUSION: Three-dimensional power-Doppler imaging did not have a better diagnostic performance than 2-dimensional power-Doppler imaging for the discrimination of benign from malignant complex adnexal masses.

Adnexa Uteri↗
PubMed2005

Predicting breast cancer survivability: a comparison of three data mining methods.

OBJECTIVE: The prediction of breast cancer survivability has been a challenging research problem for many researchers. Since the early dates of the related research, much advancement has been recorded in several related fields. For instance, thanks to innovative biomedical technologies, better explanatory prognostic factors are being measured and recorded; thanks to low cost computer hardware and software technologies, high volume better quality data is being collected and stored automatically; and finally thanks to better analytical methods, those voluminous data is being processed effectively and efficiently. Therefore, the main objective of this manuscript is to report on a research project where we took advantage of those available technological advancements to develop prediction models for breast cancer survivability. METHODS AND MATERIAL: We used two popular data mining algorithms (artificial neural networks and decision trees) along with a most commonly used statistical method (logistic regression) to develop the prediction models using a large dataset (more than 200,000 cases). We also used 10-fold cross-validation methods to measure the unbiased estimate of the three prediction models for performance comparison purposes. RESULTS: The results indicated that the decision tree (C5) is the best predictor with 93.6% accuracy on the holdout sample (this prediction accuracy is better than any reported in the literature), artificial neural networks came out to be the second with 91.2% accuracy and the logistic regression models came out to be the worst of the three with 89.2% accuracy. CONCLUSION: The comparative study of multiple prediction models for breast cancer survivability using a large dataset along with a 10-fold cross-validation provided us with an insight into the relative prediction ability of different data mining methods. Using sensitivity analysis on neural network models provided us with the prioritized importance of the prognostic factors used in the study.

Breast Neoplasms↗
PubMed2006

Predicting survival in patients requiring renal replacement therapy after cardiac surgery.

BACKGROUND: We sought to develop and internally validate a prediction score for all-cause in-hospital mortality in patients who have acute renal failure and require renal replacement therapy after cardiac surgery. METHODS: From January 1992 to July 2001, 136 of 14,000 patients (0.9%) had acute renal failure requiring renal replacement therapy after cardiac surgery. Multivariate logistic regression analysis, based on pre-renal replacement therapy variables, was used to construct a predictive score for all causes of in-hospital mortality. Subsequently, the score was validated in 27 patients who underwent surgery between August 2001 and March 2003. RESULTS: In-hospital mortality was 58% (79 of 136). From the logistic regression model, we assigned a score (range, 0 to 6) based on the presence of independent predictors of operative mortality (preoperative creatinine < or = 1.5 mg/dL [odds ratio (OR) = 5.0], hypertension [OR = 4.4], predialysis coma [OR = 9.6], sepsis [OR = 6.4], and total bilirubin > or = 2 mg/dL [OR = 5.6]). Higher scores strongly predicted mortality: patients who scored 3 or higher before the initiation dialysis (n = 54), had a mortality rate of 94% (51 of 54). In contrast, patients who scored 1 or less on this scale (n = 36), had a mortality of 16% (6 of 36). In the validation cohort, the sensitivity of the new score at the cutoff of 2 or fewer points versus 3 or more points was 0.71, the specificity was 0.90, the positive predictive value was 0.92, and the negative predictive value was 0.64. CONCLUSIONS: The prediction score represents a simple and accurate tool for predicting in-hospital mortality associated with renal replacement therapy for cardiac surgery patients before the institution of this resource-intensive treatment.

Acute Kidney Injury↗
PubMed2006

How reliable is the structural prediction of IgE-binding epitopes of allergens? The case study of plant lipid transfer proteins.

The linear IgE-binding epitopes of non-specific lipid transfer proteins (nsLTP) from plants were predicted using a combination of predictive tools including (1) the hydropathic profiles based on different scales of hydrophilicity, flexibility and exposure to the solvent, (2) the hydrophobic cluster analysis plots, (3) the occurrence of charged residues in the predicted amino acid sequence stretches and, (4) the exposition of the predicted linear IgE-binding epitopes checked on the three-dimensional models built for the nsLTP. A reliable prediction was obtained for nsLTP as compared with the previously characterized IgE-binding epitopes of various proteins. A consensual IgE-binding epitope occurring in other plant nsLTP and responsible for some IgE-binding cross-reactivity among fruit nsLTP has been identified and characterized. Despite some discrepancies, a fairly good prediction resulted in applying our combination of predictive methods to longer nsLTP or plant profilins.

Amino Acid Sequence↗
PubMed2004

Does follow-up duration influence the ultrasound and DXA prediction of hip fracture? The EPIDOS prospective study.

While the potential of quantitative ultrasound (QUS) in the management of osteoporosis has been accepted, its interaction with follow-up time has never been investigated. The aim of our study is to prospectively evaluate the influence of follow-up time on the prediction of hip fracture by ultrasound parameters in the elderly as compared to bone mineral density (BMD) and to establish a long-term fracture prediction model. In the multicenter prospective study EPIDOS, 5898 Caucasian healthy women, aged 75 and over, had femoral dual-energy X-ray absorptiometry (DXA) and heel ultrasound measurements at baseline. A survey of fracture occurrence was conducted every 4 months. Statistical analyses were performed for three different average lengths of follow-up, namely, 1.5, 2.5 and 3.5 years. Relative risks per standard deviation decrease (RR) and the area under the receiver operating characteristic (AUC) curves were given. Estimates of the long-term hip fracture prediction by DXA and QUS were extrapolated. During an average of 3.5 years follow-up, 227 women sustained their first non-traumatic hip fracture. For the three categories of follow-up, low values of both calcaneal ultrasound and hip BMD were associated with a significant increased risk of hip fracture [e.g. ultrasound Stiffness index RR = 2.8 (2.1-3.8), 2.1 (1.7-2.6) and 1.9 (1.7-2.3) for 1.5, 2.5 and 3.5 years of follow-up, respectively]. The combination of femoral neck BMD with the Stiffness showed an improvement of the hip fracture prediction model. Using extrapolation, the prediction of hip fracture by the Stiffness remained significant up to 7.5 years [RR = 1.2 (1.03-1.41)], whereas the limit of significance was reached at 10 years for the femoral neck BMD [RR = 1.25 (1.04-1.52)]. Our results indicate that the Stiffness tends to be the best short- and long-term predictor of hip fracture among ultrasound parameters. This paper provides additional information on the long-term prediction of hip fracture, which has always been an important issue in routine clinical practice as it influences the management of the disease. Our model should give a relatively good estimation of the fracture risk prediction at 5 years with the ultrasound and 10 years for the femoral neck BMD.

Absorptiometry, Photon↗
PubMed2005

Modulation of early auditory processing by visually based sound prediction.

Brain activity was measured by magnetoencephalography (MEG) to investigate whether the early auditory system can detect changes in audio-visual patterns when the visual part is presented earlier. We hypothesized that a template underlying the mismatch field (MMF) phenomenon, which is usually formed by past sound regularities, is also used in visually based sound prediction. Activity similar to the MMF may be elicited by comparing an incoming sound with the template. The stimulus was modeled after a keyboard: an animation in which one of two keys was depressed was accompanied by either a lower or higher tone. Congruent audio-visual pairs were designed to be frequent and incongruent pairs to be infrequent. Subjects were instructed to predict an incoming sound based on key movement in two sets of trials (prediction condition), whereas they were instructed not to do so in the other two sets (non-prediction condition). For each condition, the movement took 50 ms in one set (Delta = 50 ms) and 300 ms in the other (Delta = 300 ms) to reach the bottom, at which time a tone was delivered. As a result, only under the prediction condition with Delta = 300 ms was additional activity for incongruent pairs observed bilaterally in the supratemporal area within 100-200 ms of the auditory stimulus onset; this activity had spatio-temporal properties similar to those of MMF. We concluded that a template is created by the visually based sound prediction only after the visual discriminative and sound prediction processes have already been performed.

Acoustic Stimulation↗
PubMed2005

Predicting the behaviour of proteins in hydrophobic interaction chromatography. 2. Using a statistical description of their surface amino acid distribution.

This paper focuses on the prediction of the dimensionless retention time (DRT) of proteins in hydrophobic interaction chromatography (HIC) by means of mathematical models based on the statistical description of the amino acid surface distribution. Previous models characterises the protein surface as a whole. However, most of the time it is not the whole protein but some of its specific regions that interact with the environment. It seems much more natural to use local measurements of the characteristics of the surface. Therefore, the statistical characterisation of the distribution of an amino acid property on the protein surface was carried out from the systematic calculation of the local average of this property in a neighbourhood placed sequentially on each of the amino acids on the protein surface. This process allowed us to characterise the distribution of this property quantitatively using three main statistics: average, standard deviation and maximum. In particular, if the property considered is a hydrophobicity scale, these statistics allowed us to characterise the average hydrophobicity and the hydrophobic content of the most hydrophobic cluster or hotspot, as well as the heterogeneity of the hydrophobicity distribution on the protein surface. We tested the performance of the DRT predictive models based on these statistics on a set of 15 proteins. We obtained better predictive results with respect to the models previously reported. The best predictive model was a linear model based on the maximum. This statistic was calculated using an index of the mobilities of amino acids in chromatography. The predictive performance of this model (measured as the Jack Knife MSE) was 26.9% better than those obtained by the best model which does not consider the amino acid distribution and 19.5% better than the model based on the hydrophobic imbalance (HI). In addition, the best performance was obtained by a linear multivariable model based on the HI and the maximum. The difference between the experimental data and the prediction carried out by this model was smaller than those observed previously. In fact, this model obtained better predictive capacities than a previous linear multivariable model decreasing the Jack Knife MSE in 8.7%. In addition, this model allowed us to diminish the number of variables required, increasing, in this way, the degrees of freedom of the model.

Amino Acids↗