Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues

Desu Chen, Sumona Sarkar, Julián Candia, Stephen J. Florczyk, Subhadip Bodhak, Meghan K. Driscoll, Carl G. Simon, Joy P. Dunkers, Wolfgang Losert

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Cell morphology has been identified as a potential indicator of stem cell response to biomaterials. However, determination of cell shape phenotype in biomaterials is complicated by heterogeneous cell populations, microenvironment heterogeneity, and multi-parametric definitions of cell morphology. To associate cell morphology with cell-material interactions, we developed a shape phenotyping framework based on support vector machines. A feature selection procedure was implemented to select the most significant combination of cell shape metrics to build classifiers with both accuracy and stability to identify and predict microenvironment-driven morphological differences in heterogeneous cell populations. The analysis was conducted at a multi-cell level, where a “supercell” method used average shape measurements of small groups of single cells to account for heterogeneous populations and microenvironment. A subsampling validation algorithm revealed the range of supercell sizes and sample sizes needed for classifier stability and generalization capability. As an example, the responses of human bone marrow stromal cells (hBMSCs) to fibrous vs flat microenvironments were compared on day 1. Our analysis showed that 57 cells (grouped into supercells of size 4) are the minimum needed for phenotyping. The analysis identified that a combination of minor axis length, solidity, and mean negative curvature were the strongest early shape-based indicator of hBMSCs response to fibrous microenvironment.

Original languageEnglish (US)
Pages (from-to)104-118
Number of pages15
JournalBiomaterials
Volume104
DOIs
StatePublished - Oct 1 2016
Externally publishedYes

Fingerprint

Cell Shape
Cues
Learning systems
Biocompatible Materials
Phenotype
Biomaterials
Bone
Classifiers
Cells
Stem cells
Support vector machines
Feature extraction
Mesenchymal Stromal Cells
Cellular Microenvironment
Population Characteristics
Machine Learning
Cell Communication
Sample Size
Population
Stem Cells

Keywords

  • Cell morphology
  • Fibrous substrates
  • Machine learning
  • Stem cell
  • Supercell

ASJC Scopus subject areas

  • Bioengineering
  • Ceramics and Composites
  • Biophysics
  • Biomaterials
  • Mechanics of Materials

Cite this

Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues. / Chen, Desu; Sarkar, Sumona; Candia, Julián; Florczyk, Stephen J.; Bodhak, Subhadip; Driscoll, Meghan K.; Simon, Carl G.; Dunkers, Joy P.; Losert, Wolfgang.

In: Biomaterials, Vol. 104, 01.10.2016, p. 104-118.

Research output: Contribution to journalArticle

Chen, D, Sarkar, S, Candia, J, Florczyk, SJ, Bodhak, S, Driscoll, MK, Simon, CG, Dunkers, JP & Losert, W 2016, 'Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues', Biomaterials, vol. 104, pp. 104-118. https://doi.org/10.1016/j.biomaterials.2016.06.040
Chen, Desu ; Sarkar, Sumona ; Candia, Julián ; Florczyk, Stephen J. ; Bodhak, Subhadip ; Driscoll, Meghan K. ; Simon, Carl G. ; Dunkers, Joy P. ; Losert, Wolfgang. / Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues. In: Biomaterials. 2016 ; Vol. 104. pp. 104-118.
@article{27b0af68d8a9415dad3302c5e51ce866,
title = "Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues",
abstract = "Cell morphology has been identified as a potential indicator of stem cell response to biomaterials. However, determination of cell shape phenotype in biomaterials is complicated by heterogeneous cell populations, microenvironment heterogeneity, and multi-parametric definitions of cell morphology. To associate cell morphology with cell-material interactions, we developed a shape phenotyping framework based on support vector machines. A feature selection procedure was implemented to select the most significant combination of cell shape metrics to build classifiers with both accuracy and stability to identify and predict microenvironment-driven morphological differences in heterogeneous cell populations. The analysis was conducted at a multi-cell level, where a “supercell” method used average shape measurements of small groups of single cells to account for heterogeneous populations and microenvironment. A subsampling validation algorithm revealed the range of supercell sizes and sample sizes needed for classifier stability and generalization capability. As an example, the responses of human bone marrow stromal cells (hBMSCs) to fibrous vs flat microenvironments were compared on day 1. Our analysis showed that 57 cells (grouped into supercells of size 4) are the minimum needed for phenotyping. The analysis identified that a combination of minor axis length, solidity, and mean negative curvature were the strongest early shape-based indicator of hBMSCs response to fibrous microenvironment.",
keywords = "Cell morphology, Fibrous substrates, Machine learning, Stem cell, Supercell",
author = "Desu Chen and Sumona Sarkar and Juli{\'a}n Candia and Florczyk, {Stephen J.} and Subhadip Bodhak and Driscoll, {Meghan K.} and Simon, {Carl G.} and Dunkers, {Joy P.} and Wolfgang Losert",
year = "2016",
month = "10",
day = "1",
doi = "10.1016/j.biomaterials.2016.06.040",
language = "English (US)",
volume = "104",
pages = "104--118",
journal = "Biomaterials",
issn = "0142-9612",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Machine learning based methodology to identify cell shape phenotypes associated with microenvironmental cues

AU - Chen, Desu

AU - Sarkar, Sumona

AU - Candia, Julián

AU - Florczyk, Stephen J.

AU - Bodhak, Subhadip

AU - Driscoll, Meghan K.

AU - Simon, Carl G.

AU - Dunkers, Joy P.

AU - Losert, Wolfgang

PY - 2016/10/1

Y1 - 2016/10/1

N2 - Cell morphology has been identified as a potential indicator of stem cell response to biomaterials. However, determination of cell shape phenotype in biomaterials is complicated by heterogeneous cell populations, microenvironment heterogeneity, and multi-parametric definitions of cell morphology. To associate cell morphology with cell-material interactions, we developed a shape phenotyping framework based on support vector machines. A feature selection procedure was implemented to select the most significant combination of cell shape metrics to build classifiers with both accuracy and stability to identify and predict microenvironment-driven morphological differences in heterogeneous cell populations. The analysis was conducted at a multi-cell level, where a “supercell” method used average shape measurements of small groups of single cells to account for heterogeneous populations and microenvironment. A subsampling validation algorithm revealed the range of supercell sizes and sample sizes needed for classifier stability and generalization capability. As an example, the responses of human bone marrow stromal cells (hBMSCs) to fibrous vs flat microenvironments were compared on day 1. Our analysis showed that 57 cells (grouped into supercells of size 4) are the minimum needed for phenotyping. The analysis identified that a combination of minor axis length, solidity, and mean negative curvature were the strongest early shape-based indicator of hBMSCs response to fibrous microenvironment.

AB - Cell morphology has been identified as a potential indicator of stem cell response to biomaterials. However, determination of cell shape phenotype in biomaterials is complicated by heterogeneous cell populations, microenvironment heterogeneity, and multi-parametric definitions of cell morphology. To associate cell morphology with cell-material interactions, we developed a shape phenotyping framework based on support vector machines. A feature selection procedure was implemented to select the most significant combination of cell shape metrics to build classifiers with both accuracy and stability to identify and predict microenvironment-driven morphological differences in heterogeneous cell populations. The analysis was conducted at a multi-cell level, where a “supercell” method used average shape measurements of small groups of single cells to account for heterogeneous populations and microenvironment. A subsampling validation algorithm revealed the range of supercell sizes and sample sizes needed for classifier stability and generalization capability. As an example, the responses of human bone marrow stromal cells (hBMSCs) to fibrous vs flat microenvironments were compared on day 1. Our analysis showed that 57 cells (grouped into supercells of size 4) are the minimum needed for phenotyping. The analysis identified that a combination of minor axis length, solidity, and mean negative curvature were the strongest early shape-based indicator of hBMSCs response to fibrous microenvironment.

KW - Cell morphology

KW - Fibrous substrates

KW - Machine learning

KW - Stem cell

KW - Supercell

UR - http://www.scopus.com/inward/record.url?scp=84978701864&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84978701864&partnerID=8YFLogxK

U2 - 10.1016/j.biomaterials.2016.06.040

DO - 10.1016/j.biomaterials.2016.06.040

M3 - Article

C2 - 27449947

AN - SCOPUS:84978701864

VL - 104

SP - 104

EP - 118

JO - Biomaterials

JF - Biomaterials

SN - 0142-9612

ER -