Content area
Abstract
A principal challenge in the analysis of tissue imaging data is cell segmentation—the task of identifying the precise boundary of every cell in an image. To address this problem we constructed TissueNet, a dataset for training segmentation models that contains more than 1 million manually labeled cells, an order of magnitude more than all previously published segmentation training datasets. We used TissueNet to train Mesmer, a deep-learning-enabled segmentation algorithm. We demonstrated that Mesmer is more accurate than previous methods, generalizes to the full diversity of tissue types and imaging platforms in TissueNet, and achieves human-level performance. Mesmer enabled the automated extraction of key cellular features, such as subcellular localization of protein signal, which was challenging with previous approaches. We then adapted Mesmer to harness cell lineage information in highly multiplexed datasets and used this enhanced version to quantify cell morphology changes during human gestation. All code, data and models are released as a community resource.
Deep learning algorithms perform as well as humans in identifying cells in tissue images.
Details
; Miller, Geneva 2 ; Moen Erick 2 ; Kong, Alex 3 ; Kagel, Adam 3 ; Dougherty, Thomas 2 ; Fullaway, Christine Camacho 3 ; McIntosh, Brianna J 4
; Leow, Ke Xuan 1 ; Schwartz, Morgan Sarah 2 ; Cole, Pavelchek 5
; Cui Sunny 6 ; Camplisson Isabella 2 ; Bar-Tal Omer 7
; Singh Jaiveer 3 ; Fong, Mara 8 ; Chaudhry Gautam 3
; Zion, Abraham 3 ; Moseley, Jackson 3 ; Warshawsky Shiri 3 ; Soon, Erin 9 ; Greenbaum, Shirley 3
; Risom Tyler 3 ; Hollmann, Travis 10
; Bendall, Sean C 3
; Leeat, Keren 7
; Graf, William 2
; Angelo, Michael 3
; Van Valen David 2
1 Stanford University, Cancer Biology Program, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956); Stanford University, Department of Pathology, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956)
2 California Institute of Technology, Division of Biology and Bioengineering, Pasadena, USA (GRID:grid.20861.3d) (ISNI:0000000107068890)
3 Stanford University, Department of Pathology, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956)
4 Stanford University, Cancer Biology Program, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956)
5 California Institute of Technology, Division of Biology and Bioengineering, Pasadena, USA (GRID:grid.20861.3d) (ISNI:0000000107068890); Washington University School of Medicine in St. Louis, St. Louis, USA (GRID:grid.4367.6) (ISNI:0000 0001 2355 7002)
6 California Institute of Technology, Department of Electrical Engineering, Pasadena, USA (GRID:grid.20861.3d) (ISNI:0000000107068890); Princeton University, Department of Computer Science, Princeton, USA (GRID:grid.16750.35) (ISNI:0000 0001 2097 5006)
7 Weizmann Institute of Science, Department of Molecular Cell Biology, Rehovot, Israel (GRID:grid.13992.30) (ISNI:0000 0004 0604 7563)
8 Stanford University, Department of Pathology, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956); Brown University, Department of Cognitive, Linguistic and Psychological Sciences, Providence, USA (GRID:grid.40263.33) (ISNI:0000 0004 1936 9094)
9 Stanford University, Department of Pathology, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956); Stanford University, Immunology Program, Stanford, USA (GRID:grid.168010.e) (ISNI:0000000419368956)
10 Memorial Sloan Kettering Cancer Center, Department of Pathology, New York, USA (GRID:grid.51462.34) (ISNI:0000 0001 2171 9952)