Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat
Sign Up
Cited by 1Digital Holography and MicroscopyCell Image Analysis TechniquesDigital Imaging for Blood Diseases

The DOI matches its Crossref record; 59 of the 59 references it cites were found in a registry or an index.

Cell-TIMP: Cellular Trajectory Inference Based on Morphological Parameters

Open access
Raj, Piyush, Gupta, Himanshu, Anantha, Pooja, Barman, Ishan·2025·Nano Letters
DOI 10.1021/acs.nanolett.5c01009PMC PMC11071304

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Cellular morphology, shaped by various genetic and environmental influences, is pivotal to studying experimental cell biology, necessitating precise measurement and analysis techniques. Traditional approaches, which rely on geometric metrics derived from stained images, encounter obstacles stemming from both the imaging and analytical domains. Staining processes can disrupt the cell's natural state and diminish accuracy due to photobleaching, while conventional analysis techniques, which categorize cells based on shape to discern pathophysiological conditions, often fail to capture the continuous and asynchronous nature of biological processes such as cell differentiation, immune responses, and cancer progression. In this work, we propose the use of quantitative phase imaging for morphological assessment due to its label-free nature. For analysis, we repurposed the genomic analysis toolbox to perform trajectory inference analysis purely based on morphology information. We applied the developed framework to study the progression of leukemia and breast cancer metastasis. Our approach revealed a clear pattern of morphological evolution tied to the diseases' advancement, highlighting the efficacy of our method in identifying functionally significant shape changes where conventional techniques falter. This advancement offers a fresh perspective on analyzing cellular morphology and holds significant potential for the broader research community, enabling a deeper understanding of complex biological dynamics.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1021/acs.nanolett.5c01009

Cites

59 of 59 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1: Schematic of the analytical pipeline for Cell-TIMP. (a) QPI cell images are taken from the microscope. The cell images are then segmented, and a cell mask is obtained. The mask images are then centered and aligned along its major axis. Thereafter, Zernike polynomial calculation is done to obtain Zernike moments for each basis function. A reconstructed figure shows that N = 30 is sufficient for most cell morphology application. (b) Once we have converted cell morphology in a one-dimensional vector, we implement TI analytical pipeline on the data. The method is rooted in partition-based graph abstraction algorithm (PAGA) which consists of partitioning a K-nearest neighbors (K-NN) graph into a given number of clusters and linking these clusters to form trajectories. The partitioning of graph was done with leiden clustering which allows user to avoid guessing the number of clusters required and UMAP was used for low-dimensional visualization purpose. (c) A schematic depicting the pathway for Waddington landscape plot.
Figure 1: Schematic of the analytical pipeline for Cell-TIMP. (a) QPI cell images are taken from the microscope. The cell images are then segmented, and a cell mask is obtained. The mask images are then centered and aligned along its major axis. Thereafter, Zernike polynomial calculation is done to obtain Zernike moments for each basis function. A reconstructed figure shows that N = 30 is sufficient for most cell morphology application. (b) Once we have converted cell morphology in a one-dimensional vector, we implement TI analytical pipeline on the data. The method is rooted in partition-based graph abstraction algorithm (PAGA) which consists of partitioning a K-nearest neighbors (K-NN) graph into a given number of clusters and linking these clusters to form trajectories. The partitioning of graph was done with leiden clustering which allows user to avoid guessing the number of clusters required and UMAP was used for low-dimensional visualization purpose. (c) A schematic depicting the pathway for Waddington landscape plot.CC BY-NC-ND · PubMed Central
Figure 2: Morphological analysis on computer generated shapes. (a) Sample shapes created by 3 and 4 points bezier curves. (b) Spectra plot of mean Zernike moments for identified clusters for different shapes. (c) Scatter plot on a UMAP representation highlighting different identified clusters through Leiden clustering. (d) Reconstructed shape from average Zernike moment value of identified cluster.
Figure 2: Morphological analysis on computer generated shapes. (a) Sample shapes created by 3 and 4 points bezier curves. (b) Spectra plot of mean Zernike moments for identified clusters for different shapes. (c) Scatter plot on a UMAP representation highlighting different identified clusters through Leiden clustering. (d) Reconstructed shape from average Zernike moment value of identified cluster.CC BY-NC-ND · PubMed Central
Figure 3: Morphological analysis on Swedish leaf dataset. (a) UMAP plot with actual label after converting different leaf shapes into Zernike polynomials. (b) UMAP plot with unsupervised leiden clusters. (c) Plot showing sample leaf image for each label and the average reconstructed shapes of dominant leiden clusters. Below each figure, we have a colored stacked bar plot for percentage contribution of each leiden cluster constituting the given leaf. (d) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the image in the dataset. The solid black line along the y-axis is the margin for each cluster.
Figure 3: Morphological analysis on Swedish leaf dataset. (a) UMAP plot with actual label after converting different leaf shapes into Zernike polynomials. (b) UMAP plot with unsupervised leiden clusters. (c) Plot showing sample leaf image for each label and the average reconstructed shapes of dominant leiden clusters. Below each figure, we have a colored stacked bar plot for percentage contribution of each leiden cluster constituting the given leaf. (d) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the image in the dataset. The solid black line along the y-axis is the margin for each cluster.CC BY-NC-ND · PubMed Central
Figure 4: Morphological analysis on leukemia dataset. (a) Sample images of healthy B-cells (Control), early (REH, RS4;11) and late (BALL-1, MN60) stages of malignancy. (b) UMAP plot with actual label after converting cell shape into Zernike polynomials. (c) UMAP plot with unsupervised clustering labels. (d) Reconstructed shape from average value of computed Zernike polynomials from each cluster. (e) Leiden cluster distribution for each supervised label showing the majority of clusters along the differentiation timepoint align well with the pseudotime axis. (f) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters that are aligned on a pseudotime axis. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the images in the dataset. The solid black line along the y-axis is the margin for each cluster.
Figure 4: Morphological analysis on leukemia dataset. (a) Sample images of healthy B-cells (Control), early (REH, RS4;11) and late (BALL-1, MN60) stages of malignancy. (b) UMAP plot with actual label after converting cell shape into Zernike polynomials. (c) UMAP plot with unsupervised clustering labels. (d) Reconstructed shape from average value of computed Zernike polynomials from each cluster. (e) Leiden cluster distribution for each supervised label showing the majority of clusters along the differentiation timepoint align well with the pseudotime axis. (f) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters that are aligned on a pseudotime axis. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the images in the dataset. The solid black line along the y-axis is the margin for each cluster.CC BY-NC-ND · PubMed Central
Figure 5: Morphological analysis on breast cancer dataset. (a) Sample QPI images of Parental 231 (P231), Circulating tumor cells (CTC) and Lung Metastasis (LM) cells. (b) UMAP plot with actual label after converting cell shape into Zernike polynomials. (c) UMAP plot with unsupervised clustering labels. (d) Reconstructed shape from average value of computed Zernike polynomials from each cluster. (e) Leiden cluster distribution for each supervised label showing most clusters along the differentiation timepoint align well with the pseudotime axis. (f) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters that are aligned on a pseudotime axis. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the images in the dataset. The solid black line along the y-axis is the margin for each cluster.
Figure 5: Morphological analysis on breast cancer dataset. (a) Sample QPI images of Parental 231 (P231), Circulating tumor cells (CTC) and Lung Metastasis (LM) cells. (b) UMAP plot with actual label after converting cell shape into Zernike polynomials. (c) UMAP plot with unsupervised clustering labels. (d) Reconstructed shape from average value of computed Zernike polynomials from each cluster. (e) Leiden cluster distribution for each supervised label showing most clusters along the differentiation timepoint align well with the pseudotime axis. (f) Heatmap of top 5 variables from each cluster which were responsible for identification of clusters that are aligned on a pseudotime axis. The variable number on the left stands for Zernike polynomial index and each row corresponds to the value of Zernike moment for the given polynomial for all the images in the dataset. The solid black line along the y-axis is the margin for each cluster.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 1

  1. 10.3928/1081-597X-20020901-30 · doi:10.3928/1081-597x-20020901-30

Metadata

Authors
Raj, P., Gupta, H., Anantha, P., & Barman, I.
Affiliations
  • Johns Hopkins University (US)
  • Örebro University (SE)
Corresponding author
Barman, Ishan
ORCID
Venue
Nano Letters
Publisher
American Chemical Society (ACS)
Published
2025-05-03
Volume
25
Issue
19
Pages
7845-7852
ISSN
  • 1530-6984 (ISSN-L)
  • 1530-6992
DOI
10.1021/acs.nanolett.5c01009
PMC
PMC11071304
OpenAlex
W4410057187
Cited by
1
References
59
FWCI
0.00
Type
journal article
Language
en
Open access
Licence
Locations
  • Nano Letters (publishedVersion)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by-nc-nd)
Funders
  • Air Force Office of Scientific Research (FA955022-10334)
  • National Institute of General Medical Sciences (1R35GM149272)
Topics
  • Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Digital Holography and Microscopy
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Biophysics › Cell Image Analysis Techniques
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Imaging for Blood Diseases
Keywords
cell morphology, quantitative phase imaging, label-free imaging, morphometrics, leukemia progression, breast cancer metastasis, disease progression, cell shape changes

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.