Authors

Anthony Culos, Stanford University School of Medicine
Amy S. Tsai, Stanford University School of Medicine
Natalie Stanley, Stanford University School of Medicine
Martin Becker, Stanford University School of Medicine
Mohammad S. Ghaemi, Stanford University School of Medicine
David R. McIlwain, Stanford University School of Medicine
Ramin Fallahzadeh, Stanford University School of Medicine
Athena Tanada, Stanford University School of Medicine
Huda Nassar, Stanford University School of Medicine
Camilo Espinosa, Stanford University School of Medicine
Maria Xenochristou, Stanford University School of Medicine
Edward Ganio, Stanford University School of Medicine
Laura Peterson, Stanford University School of Medicine
Xiaoyuan Han, University of the PacificFollow
Ina A. Stelzer, Stanford University School of Medicine
Kazuo Ando, Stanford University School of Medicine
Dyani Gaudilliere, Stanford University School of Medicine
Thanaphong Phongpreecha, Stanford University School of Medicine
Ivana Marić, Stanford University School of Medicine
Alan L. Chang, Stanford University School of Medicine
Gary M. Shaw, Stanford University School of Medicine
David K. Stevenson, Stanford University School of Medicine
Sean Bendall, Stanford University School of Medicine
Kara L. Davis, Stanford University School of Medicine
Wendy Fantl, Stanford University School of Medicine
Garry P. Nolan, Stanford University School of Medicine
Trevor Hastie, Stanford University
Robert Tibshirani, Stanford University
Martin S. Angst, Stanford University School of Medicine

Document Type

Article

Publication Title

Nature Machine Intelligence

ISSN

2522-5839

Volume

2

Issue

10

DOI

10.1038/s42256-020-00232-8

First Page

619

Last Page

628

Publication Date

10-1-2020

Abstract

The dense network of interconnected cellular signalling responses that are quantifiable in peripheral immune cells provides a wealth of actionable immunological insights. Although high-throughput single-cell profiling techniques, including polychromatic flow and mass cytometry, have matured to a point that enables detailed immune profiling of patients in numerous clinical settings, the limited cohort size and high dimensionality of data increase the possibility of false-positive discoveries and model overfitting. We introduce a generalizable machine learning platform, the immunological Elastic-Net (iEN), which incorporates immunological knowledge directly into the predictive models. Importantly, the algorithm maintains the exploratory nature of the high-dimensional dataset, allowing for the inclusion of immune features with strong predictive capabilities even if not consistent with prior knowledge. In three independent studies our method demonstrates improved predictions for clinically relevant outcomes from mass cytometry data generated from whole blood, as well as a large simulated dataset. The iEN is available under an open-source licence.

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Share

COinS