Combining predictors for classification using the area under the receiver operating characteristic curve

No single biomarker for cancer is considered adequately sensitive and specific for cancer screening. It is expected that the results of multiple markers will need to be combined in order to yield adequately accurate classification. Typically, the objective function that is optimized for combining markers is the likelihood function. In this article, we consider an alternative objective function-the area under the empirical receiver operating characteristic curve (AUC). We note that it yields consistent estimates of parameters in a generalized linear model for the risk score but does not require specifying the link function. Like logistic regression, it yields consistent estimation with case-control or cohort data. Simulation studies suggest that AUC-based classification scores have performance comparable with logistic likelihood-based scores when the logistic regression model holds. Analysis of data from a proteomics biomarker study shows that performance can be far superior to logistic regression derived scores when the logistic regression model does not hold. Model fitting by maximizing the AUC rather than the likelihood should be considered when the goal is to derive a marker combination score for classification or prediction.     

Human tropoelastin sequence: Dynamics of polypeptide coded by exon 6 in solution

Calorimetric studies were performed on exon 6 in powdered form and in solution [water and 2,2,2‐trifluoroethanol (TFE), a structure‐inducing solvent or cosolvent]. Dynamic dielectric spectroscopy (DDS) analyses were realized in water and 20% TFE. The major role of solvent–peptide organization is evidenced with these techniques. Calorimetric measurements reveal the structural water organization around the polypeptide as well as the presence of hydrophobic interactions in TFE solution. Dielectric measurements showed for exon 6/water a decrease of relaxations times of bulk solvent implying a faster dynamics with a slight increase of the activation entropy, suggesting that exon 6 probably creates disorder within the solvent. For TFE/water mixtures, an influence of exon 6 on its environment was seen with a relaxation associated with the exon 6/solvent interactions reinforced by storage of 72 h. Finally, exon 6/solvent interactions were clearly observed with additionof TFE. © 2009 Wiley Periodicals, Inc. Biopolymers 91: 943–952, 2009. This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley.com     

Biomarker evaluation and comparison using the controls as a reference population

The classification accuracy of a continuous marker is typically evaluated with the receiver operating characteristic (ROC) curve. In this paper, we study an alternative conceptual framework, the "percentile value." In this framework, the controls only provide a reference distribution to standardize the marker. The analysis proceeds by analyzing the standardized marker in cases. The approach is shown to be equivalent to ROC analysis. Advantages are that it provides a framework familiar to a broad spectrum of biostatisticians and it opens up avenues for new statistical techniques in biomarker evaluation. We develop several new procedures based on this framework for comparing biomarkers and biomarker performance in different populations. We develop methods that adjust such comparisons for covariates. The methods are illustrated on data from 2 cancer biomarker studies.     

Functional analysis of MUTYH mutated proteins associated with familial adenomatous polyposis

The MUTYH DNA glycosylase specifically removes adenine misincorporated by replicative polymerases opposite the oxidized purine 8-oxo-7,8-dihydroguanine (8-oxoG). A defective protein activity results in the accumulation of G > T transversions because of unrepaired 8-oxoG:A mismatches. In humans, MUTYH germline mutations are associated with a recessive form of familial adenomatous polyposis and colorectal cancer predisposition (MUTYH-associated polyposis, MAP). Here we studied the repair capacity of the MUTYH variants R171W, E466del, 137insIW, Y165C and G382D, identified in MAP patients. Following expression and purification of human proteins from a bacterial system, we investigated MUTYH incision capacity on an 8-oxoG:A substrate by standard glycosylase assays. For the first time, we employed the surface plasmon resonance (SPR) technology for real-time recording of the association/dissociation of wild-type and MUTYH variants from an 8-oxoG:A DNA substrate. When compared to the wild-type protein, R171W, E466del and Y165C variants showed a severe reduction in the binding affinity towards the substrate, while 137insIW and G382D mutants manifested only a slight decrease mainly due to a slower rate of association. This reduced binding was always associated with impairment of glycosylase activity, with adenine removal being totally abrogated in R171W, E466del and Y165C and only partially reduced in 137insIW and G382D. Our findings demonstrate that SPR analysis is suitable to identify defective enzymatic behaviour even when mutant proteins display minor alterations in substrate recognition.     

The multiple sulfatase deficiency gene encodes an essential and limiting factor for the activity of sulfatases

In multiple sulfatase deficiency (MSD), a human inherited disorder, the activities of all sulfatases are impaired due to a defect in posttranslational modification. Here we report the identification, by functional complementation using microcell-mediated chromosome transfer, of a gene that is mutated in MSD and is able to rescue the enzymatic deficiency in patients' cell lines. Functional conservation of this gene was observed among distantly related species, suggesting a critical biological role. Coexpression of SUMF1 with sulfatases results in a strikingly synergistic increase of enzymatic activity, indicating that SUMF1 is both an essential and a limiting factor for sulfatases. These data have profound implications on the feasibility of enzyme replacement therapy for eight distinct inborn errors of metabolism.     

Up-regulation of alpha-inhibin expression in the fetal ovary of estrogen-suppressed baboons is associated with impaired fetal ovarian folliculogenesis

We recently demonstrated that the number of primordial follicles was significantly reduced in the ovaries of near-term baboon fetuses deprived of estrogen in utero and restored to normal in animals administered estradiol. Although the baboon fetal ovary expressed estrogen receptors alpha and beta, the mechanism(s) of estrogen action remains to be determined. It is well established that inhibin and activins function as autocrine/paracrine factors that impact adult ovarian function. However, our understanding of the expression of these factors in the primate fetal ovary is incomplete. Therefore, we determined the expression of alpha-inhibin, activin beta(A), activin beta(B), and activin receptors in fetal ovaries obtained at mid and late gestation from untreated baboons and at late gestation from animals in which fetal estrogen levels were reduced by >95% by maternal administration of the aromatase inhibitor CGS 20267 or restored to 30% of normal by treatment with CGS 20267 and estradiol benzoate. Immunocytochemical expression of alpha-inhibin was minimal to nondetectable in fetal ovaries from untreated baboons. In contrast, in baboons depleted of estrogen, alpha-inhibin was abundantly expressed in pregranulosa cells of interfollicular nests and granulosa cells of primordial follicles. Thus, the number (mean +/- SEM) per 0.08 mm2 of fetal ovarian cells expressing alpha-inhibin, determined by image analysis, was similar at mid and late gestation and increased approximately 8-fold (P < 0.01) near term in baboons treated with CGS 20267 and was restored (P < 0.01) to normal in baboons treated with CGS 20267 plus estradiol. Activin beta(A) was detected in oocytes and pregranulosa cells at midgestation and in oocytes and granulosa cells of primordial follicles at late gestation. Activin beta(B) was also expressed in pregranulosa cells and granulosa cells at mid and late gestation, respectively, but was not detected in oocytes. Neither the pattern nor the apparent level of expression of activin beta(A) or beta(B) were altered in fetal ovaries of baboons treated with CGS 20267 or CGS 20267 and estrogen. Activin receptors IA, IB, IIA, and IIB were detected by Western blot analysis in fetal ovaries at mid and late gestation, and expression was not altered by treatment with CGS 20267 or CGS 20267 and estrogen. Activin receptors IB and IIA were localized to oocytes and pregranulosa cells at midgestation and to granulosa cells and oocytes of primordial follicles at late gestation. Thus, the decrease in the number of follicles in the primate fetal ovary of baboons deprived of estrogen in utero was associated with increased expression of alpha-inhibin. Therefore, we propose that estrogen regulates fetal ovarian follicular development by controlling alpha-inhibin expression and, thus, the intraovarian inhibin:activin ratio.