Establishing a many-cytokine signature via multivariate anomaly detection

Abstract

Establishing a cytokine signature associated to some medical condition is an important task in immunology. Increasingly, large numbers of cytokines are used for signatures, via lists of reference ranges for each individual cytokine or ratios of cytokines. Here we argue that this common approach has weaknesses, especially when many different cytokines are analysed. Instead, we propose that establishing signatures can be framed as a multivariate anomaly detection problem, and hence exploit the many statistical methods available for this. In this framework, whether or not a given subject’s profile matches the cytokine signature of some condition is determined by whether or not the profile is typical of reference samples of that condition, as judged by an anomaly detection algorithm. We examine previously published cytokine data sets associated to pregnancy complications, brain tumours, and rheumatoid arthritis, as well as normal healthy control samples, and test the performance of a range of anomaly detection algorithms on these data, identifying the best performing methods. Finally, we suggest that this anomaly detection approach could be adopted more widely for general multi-biomarker signatures.

<p>CytokineExplore: An Online Tool for Statistical Analysis of Cytokine Concentration Datasets</p>

“…Such information may be valuable for further pathophysiological, diagnostic or even immunomodulation studies and may also suggest a pattern that can be used as a predictive tool to follow the progression and prognosis of the disease. 3,10,17,[31][32][33]39 Some limitations of the tool are that it performs only a battery of predetermined functions, whereas users may require more bespoke operations for some given investigation. If these are required, then the user would need to implement these themselves.…”

“…We hope that the introduction of CytokineExplore will motivate the development of other similar tools for other computational tasks in immunology and biology. For example, we recently showed 33 that sophisticated multivariate tools for outlier analysis could be directly applied to cytokine data, detecting anomalous cytokine profiles, and developing signatures. Nonetheless, the potential inability to implement such tools would inhibit application of these findings.…”

“…[5][6][7][8] It has also been suggested that cytokines can serve as disease markers or may be used to follow up on progression of the disease/therapy. 2,5,[8][9][10] Accordingly, there is enormous number of reports and studies that focuses on the level of single cytokines in a wide range of diseases. 1,2,5,6,[11][12][13][14] Th1 and Th2, the two major subsets of CD4+T helper cells have different patterns of cytokine production and different roles in immune responses.…”

“…The number of studies that have reported levels of individual cytokines in different conditions is enormous, but may be insufficient unless data are further analysed to describe the full architecture as a "signed" by the profiles of the several cytokines produced simultaneously. 10,[30][31][32][33] Until recently, measuring multiple cytokines was both expensive and required analyzing relatively large volumes of biological fluid. Nowadays, however, multiplex arrays have made measuring tens and even hundreds of cytokines at the same time less expensive and easier.…”

Multivariate Comparison of Cytokine Profiles for Normal- and Low-Bone-Density Subjects

“…Given that cytokines form a network of interacting entities and that a single cytokine or a ratio of two may not provide sufficient information about the overall immune reactivity, it is of interest to study the combined levels of several cytokines as this may provide a better picture of immune reactivity. Indeed, we and others have advocated multivariate and advanced statistical methods in cytokine data analysis studies [16][17][18].…”