ROC (receiver operating characteristic) curve: Principles and application in biology

Laboratoire de biochimie, toxicologie cliniques, Hôpital d'Instruction des Armées du Val-de-Grâce, Paris, France.
Annales de biologie clinique (Impact Factor: 0.28). 03/2005; 63(2):145-54.
Source: PubMed


Laboratory test's diagnostic performances are generally estimated by means of their sensibility, specificity and positive and negative predictive values. Unfortunately, these indices reflect only imperfectly the capacity of a test to correctly classify subjects into clinically relevant subgroups. The appeal to ROC (receiver operating characteristic) curve appears as a tool of choice for this evaluation. Used in the medical domain since the 60s, ROC curve is a graphic representation of the relation existing between the sensibility and the specificity of a test, calculated for all possible cut-off. It allows the determination and the comparison of the diagnostic performances of several tests. It is also used to consider the optimal cut-off of a test, by taking into account epidemiological and medical - economic data of the disease. Used in numerous medical domains, this statistical tool is easily accessible thanks to the development of computer softwares. This article exposes the principles of construction and exploitation of a ROC curve.

18 Reads
  • Source
    • "The peak of the resulting profile is considered to correspond to the location of the B-cell epitope for the protein concerned . Unfortunately, to one's disappointment, a recent study by Blythe and Flower (2005) led to the conclusion that ''single-scale amino acid propensity profiles cannot be used to predict epitope location reliably'', as reflected by the fact that even the best amino acid propensity scales could only yield a success rate marginally better than that by randomly using ROC (receiver operating characteristics) plot (Delacour et al., 2005). As the data of linear B-cell epitopes are accumulating, it is possible to increase the prediction quality by using machine learning approaches (Sollner, 2006; Sollner and Mayer, 2006), and the prediction accuracy has been improved by so doing. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Identification of antigenic sites on proteins is of vital importance for developing synthetic peptide vaccines, immunodiagnostic tests and antibody production. Currently, most of the prediction algorithms rely on amino acid propensity scales using a sliding window approach. These methods are oversimplified and yield poor predicted results in practice. In this paper, a novel scale, called the amino acid pair (AAP) antigenicity scale, is proposed that is based on the finding that B-cell epitopes favor particular AAPs. It is demonstrated that, using SVM (support vector machine) classifier, the AAP antigenicity scale approach has much better performance than the existing scales based on the single amino acid propensity. The AAP antigenicity scale can reflect some special sequence-coupled feature in the B-cell epitopes, which is the essence why the new approach is superior to the existing ones. It is anticipated that with the continuous increase of the known epitope data, the power of the AAP antigenicity scale approach will be further enhanced.
    Amino Acids 10/2007; 33(3):423-8. DOI:10.1007/s00726-006-0485-9 · 3.29 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The ability to predict antigenic sites on proteins is crucial for the production of synthetic peptide vaccines and synthetic peptide probes of antibody structure. Large number of amino acid propensity scales based on various properties of the antigenic sites like hydrophilicity, flexibility/mobility, turns and bends have been proposed and tested previously. However these methods are not very accurate in predicting epitopes and non-epitope regions. We propose algorithms that combine 14 best performing individual propensity scales and give better prediction accuracy as compared to individual scales.
    Computational Systems Bioinformatics Conference, 2005. Workshops and Poster Abstracts. IEEE; 09/2005
  • [Show abstract] [Hide abstract]
    ABSTRACT: This study was undertaken to quantify tissue factor (TF) and vascular endothelial growth factor (VEGF) in colorectal cancer and to evaluate their possible relationship with recurrence. TF and VEGF were measured by enzyme-linked immunosorbent assay in surgical tumour specimens and normal mucosa from 50 consecutive patients with colorectal cancer who were followed up for 3 years for the assessment of disease recurrence. TF and VEGF antigens were detected in all tumour samples. VEGF, but not TF, was much higher in tumour than in normal mucosa (P < 0.0001), as also confirmed by measurement of specific mRNAs. There was a strong correlation between TF and VEGF antigens (P < 0.0005) in tumour tissue but not in normal mucosa. Neither protein was related to tumour stage, grade or size. Local or distant recurrence was statistically related to pTNM stage. High VEGF, but not TF, levels in tumour extracts were associated with an increased risk of recurrence both by univariate (RR, 4.00, 95% CI: 1.45-11.0) and multivariate analyses (RR, 3.65, 95% CI: 1.33-10.0). These findings suggest that VEGF content in colorectal cancer is an independent risk factor for tumour recurrence and might help select patients who might benefit from adjuvant therapy.
    Colorectal Disease 03/2007; 9(2):133-8. DOI:10.1111/j.1463-1318.2006.01158.x · 2.35 Impact Factor
Show more