Classifying Antibodies Using Flow Cytometry Data: Class Prediction and Class Discovery

• Published in: Biometrical journal Vol. 47; no. 5; pp. 740 - 754
• Main Authors: Salganik, M. P., Milford, E. L., Hardie, D. L., Shaw, S., Wand, M. P.
• Format: Journal Article
• Language: English
• Published: Berlin WILEY-VCH Verlag 01-10-2005; WILEY‐VCH Verlag; Wiley-VCH
• Subjects: Antibodies, Monoclonal - classification; Antibody Specificity; Applications; Applied sciences; Biology, psychology, social sciences; Cell Line, Tumor; Class discovery; Class prediction; Classification; Computer science; control theory; systems; Control theory. Systems; Dissimilarity measure; Double-Blind Method; Exact sciences and technology; Flow cytometry; Flow Cytometry - methods; Flow Cytometry - statistics & numerical data; Fluorescent Dyes; Humans; Kernel smoothing; Mathematics; Medical sciences; Monoclonal antibodies; Predictive Value of Tests; Probability and statistics; Probability theory and stochastic processes; Sciences and techniques of general use; SiZer; Special processes (renewal theory, markov renewal processes, semi-markov processes, statistical mechanics type models, applications); Statistics; Supervised learning; System theory; Unsupervised learning; Biometrics; Flow cytometry; Process computer; Class discovery; Monoclonal antibody; Data flow; Immunology; SiZer; Classification; Computer; Smoothing methods; Dissimilarity measure; Life science; Class prediction; Interpretation; Unsupervised learning; Protein; Kernel method; Medicine; Statistical method; Cell population; Supervised learning; Monoclonal antibodies; Random set; Planning; Kernel smoothing; Algorithm analysis
• ISSN: 0323-3847; 1521-4036
• DOI: 10.1002/bimj.200310142

Classifying monoclonal antibodies, based on the similarity of their binding to the proteins (antigens) on the surface of blood cells, is essential for progress in immunology, hematology and clinical medicine. The collaborative efforts of researchers from many countries have led to the classification of thousands of antibodies into 247 clusters of differentiation (CD). Classification is based on flow cytometry and biochemical data. In preliminary classifications of antibodies based on flow cytometry data, the object requiring classification (an antibody) is described by a set of random samples from unknown densities of fluorescence intensity. An individual sample is collected in the experiment, where a population of cells of a certain type is stained by the identical fluorescently marked replicates of the antibody of interest. Samples are collected for multiple cell types. The classification problems of interest include identifying new CDs (class discovery or unsupervised learning) and assigning new antibodies to the known CD clusters (class prediction or supervised learning).
These problems have attracted limited attention from statisticians. We recommend a novel approach to the classification process in which a computer algorithm suggests to the analyst the subset of the “most appropriate” classifications of an antibody in class prediction problems or the “most similar” pairs/groups of antibodies in class discovery problems. The suggested algorithm speeds up the analysis of a flow cytometry data by a factor 10–20. This allows the analyst to focus on the interpretation of the automatically suggested preliminary classification solutions and on planning the subsequent biochemical experiments. (© 2005 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)