Analysis of the Human Protein Atlas Weakly Supervised Single-Cell Classification competition

Le, Trang; Winsnes, Casper F.; Axelsson, Ulrika; Xu, Hao; Kaimal, Jayasankar Mohanakrishnan; Mahdessian, Diana; Dai, Shubin; Makarov, Ilya S; Ostankovich, Vladislav; Xu, Yang; Benhamou, Eric; Henkel, Christof; Solovyev, Roman; Banić, Nikola; Bošnjak, Vito; Bošnjak, Ana; Miličević, Andrija; Ouyang, Wei; Lundberg, Emma

doi:10.1038/s41592-022-01606-z

Cited by 21 publications

(26 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We believe that other research questions could benefit from this type of screening. As an example, the Human Protein Atlas Image Classification competition (Ouyang et al, 2019;Le et al, 2022) managed to classify multiple organelles of individual cells in fluorescence microscopy. Such machine-learning models could be used to find rare events or particularly interesting phenotypes.…”

Section: Discussionmentioning

confidence: 99%

CLEMSite, a software for automated phenotypic screens using light microscopy and FIB-SEM

Lleti

Steyer

Schieber

et al. 2022

Journal of Cell Biology

View full text Add to dashboard Cite

In recent years, Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) has emerged as a flexible method that enables semi-automated volume ultrastructural imaging. We present a toolset for adherent cells that enables tracking and finding cells, previously identified in light microscopy (LM), in the FIB-SEM, along with the automatic acquisition of high-resolution volume datasets. We detect the underlying grid pattern in both modalities (LM and EM), to identify common reference points. A combination of computer vision techniques enables complete automation of the workflow. This includes setting the coincidence point of both ion and electron beams, automated evaluation of the image quality and constantly tracking the sample position with the microscope’s field of view reducing or even eliminating operator supervision. We show the ability to target the regions of interest in EM within 5 µm accuracy while iterating between different targets and implementing unattended data acquisition. Our results demonstrate that executing volume acquisition in multiple locations autonomously is possible in EM.

show abstract

Section: Discussionmentioning

confidence: 99%

CLEMSite, a software for automated phenotypic screens using light microscopy and FIB-SEM

Lleti

Steyer

Schieber

et al. 2022

Journal of Cell Biology

View full text Add to dashboard Cite

show abstract

“…As a result, different parts of our system can be employed as stand-alone applications in the broader context (increasing the accuracy of existing labels and reducing the labour required for manual labelling tasks). In addition, the HCPL system produces superior results to those of 11 and achieves the classification performance of 57.19% mAP in single-cell analysis.…”

Section: Introductionmentioning

confidence: 99%

“…It constitutes a core strategy of the LifeTime Initiative, a large-scale, long-term initiative to implement cell-based interceptive medicine in Europe 7 . Although machine learning (ML) has been used to describe the location of human proteins in microscope images giving summary information on an entire population of cells 8 – 10 , subcellular classification of proteins for individual cells is still an open research area with limited published work 11 and limited publicly available high-quality data 12 .…”

Section: Introductionmentioning

confidence: 99%

“…However, in contrast to ref. 11 that reviews the best approaches in the competition, our work concentrates on the challenges that single-cell classification imposes and presents a systematic and in-depth solution. As a result, different parts of our system can be employed as stand-alone applications in the broader context (increasing the accuracy of existing labels and reducing the labour required for manual labelling tasks).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Single-cell subcellular protein localisation using novel ensembles of diverse deep architectures

et al. 2023

View full text Add to dashboard Cite

Unravelling protein distributions within individual cells is vital to understanding their function and state and indispensable to developing new treatments. Here we present the Hybrid subCellular Protein Localiser (HCPL), which learns from weakly labelled data to robustly localise single-cell subcellular protein patterns. It comprises innovative DNN architectures exploiting wavelet filters and learnt parametric activations that successfully tackle drastic cell variability. HCPL features correlation-based ensembling of novel architectures that boosts performance and aids generalisation. Large-scale data annotation is made feasible by our AI-trains-AI approach, which determines the visual integrity of cells and emphasises reliable labels for efficient training. In the Human Protein Atlas context, we demonstrate that HCPL is best performing in the single-cell classification of protein localisation patterns. To better understand the inner workings of HCPL and assess its biological relevance, we analyse the contributions of each system component and dissect the emergent features from which the localisation predictions are derived.

show abstract

“…The winning model from this competition was validated and productized in the HuBMAP data portal; it is now being run on all kidney tissue data at scale. In parallel, the Human Protein Atlas (HPA) 2 conducted two Kaggle competitions [12][13][14][15] that focussed on classification of subcellular patterns in cultivated cells in microscope confocal images, engaging nearly 3,000 teams across the two competitions. In addition to the confocal images of cultivated cells, the HPA has also generated >10 million immunohistochemically stained images from all major tissues of the human body 16 .…”

Section: Introductionmentioning

confidence: 99%

Segmenting functional tissue units across human organs using community-driven development of generalizable machine learning algorithms

Jain

Godwin

Joshi

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

The development of a reference atlas of the healthy human body requires automated image segmentation of major anatomical structures across multiple organs based on spatial bioimages generated from various sources with differences in sample preparation. We present the setup and results of the 'Hacking the Human Body' machine learning algorithm development competition hosted by the Human Biomolecular Atlas (HuBMAP) and the Human Protein Atlas (HPA) teams on the Kaggle platform. We showcase how 1,175 teams from 78 countries engaged in community-driven, open-science code development that resulted in machine learning models which successfully segment anatomical structures across five organs using histology images from two consortia and that will be productized in the HuBMAP data portal to process large datasets at scale in support of Human Reference Atlas construction. We discuss the benchmark data created for the competition, major challenges faced by the participants, and the winning models and strategies.

show abstract

Analysis of the Human Protein Atlas Weakly Supervised Single-Cell Classification competition

Cited by 21 publications

References 49 publications

CLEMSite, a software for automated phenotypic screens using light microscopy and FIB-SEM

CLEMSite, a software for automated phenotypic screens using light microscopy and FIB-SEM

Single-cell subcellular protein localisation using novel ensembles of diverse deep architectures

Segmenting functional tissue units across human organs using community-driven development of generalizable machine learning algorithms

Contact Info

Product

Resources

About