WormPose: Image synthesis and convolutional networks for pose estimation in<i>C. elegans</i>

Hebert, Laetitia; Ahamed, Tosif; Costa, Á. S.; O’Shaugnessy, Liam; Stephens, Greg J.

doi:10.1101/2020.07.09.193755

Cited by 10 publications

(16 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These emergent Markov dynamics offer a promising and powerful demonstration of quantitative connections across the hierarchy of movement behavior generally exhibited by all organisms [88,112]. Particularly interesting future directions include the analysis of even longer dynamics in C. elegans [113][114][115][116], for which no canonical behavioral states have been described.…”

Section: Discussionmentioning

confidence: 99%

Maximally predictive ensemble dynamics from data

Costa

Ahamed

Jordan³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We leverage the interplay between microscopic variability and macroscopic order to connect physical descriptions across scales directly from data, without underlying equations. We reconstruct a state space by concatenating measurements in time, building a maximum entropy partition of the resulting sequences, and choosing the sequence length to maximize predictive information. Trading non-linear trajectories for linear, ensemble evolution, we analyze reconstructed dynamics through transfer operators. The evolution is parameterized by a transition time τ: capturing the source entropy rate at small τ and revealing timescale separation with collective, coherent states through the operator spectrum at larger τ. Applicable to both deterministic and stochastic systems, we illustrate our approach through the Langevin dynamics of a particle in a double-well potential and the Lorenz system. Applied to the behavior of the nematode worm C. elegans, we derive a "run-and-pirouette" navigation strategy directly from posture dynamics. We demonstrate how sequences simulated from the ensemble evolution recover effective diffusion in the worm's centroid trajectories and introduce a top-down, operator-based clustering which reveals subtle subdivisions of the "run" behavior.

show abstract

Section: Discussionmentioning

confidence: 99%

Maximally predictive ensemble dynamics from data

Costa

Ahamed

Jordan³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

Maximally predictive ensemble dynamics from data

Costa¹,

Ahamed²,

Jordan³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We leverage the interplay between microscopic variability and macroscopic order to connect physical descriptions across scales directly from data, without underlying equations. We reconstruct a state space by concatenating measurements in time, building a maximum entropy partition of the resulting sequences, and choosing the sequence length to maximize predictive information. Trading non-linear trajectories for linear, ensemble evolution, we analyze reconstructed dynamics through transfer operators. The evolution is parameterized by a transition time τ : capturing the source entropy rate at small τ and revealing timescale separation with collective, coherent states through the operator spectrum at larger τ . Applicable to both deterministic and stochastic systems, we illustrate our approach through the Langevin dynamics of a particle in a double-well potential and the Lorenz system. Applied to the behavior of the nematode worm C. elegans, we derive a "runand-pirouette" navigation strategy directly from posture dynamics. We demonstrate how sequences simulated from the ensemble evolution recover effective diffusion in the worm's centroid trajectories and introduce a top-down, operator-based clustering which reveals subtle subdivisions of the "run" behavior.

show abstract

“…elegans strains used in this work are listed in Supplementary Table 1. Worms are cultured on Nematode Growth Medium (NGM) agar at 20 ºC and fed with E. coli OP50 following standard procedures 18 .…”

Section: Worm Strainsmentioning

confidence: 99%

“…If there is sufficient spatial and temporal resolution, then body parts can be identified and tracked, the animal's pose can be estimated, and the full suite of computational ethology methods can be applied to analyse any behaviour of interest. Because of its simple morphology, detailed pose estimation is well-established for the roundworm C. elegans [5][6][7][8][9][10][11][12][13][14][15][16][17][18] and previous work has shown the usefulness of detailed behavioural phenotyping in several domains including, for example, classifying mutants 5,6,19,20 , studying chemotaxis 21 and thermotaxis 22 , quantifying escape responses [23][24][25] , and addressing basic questions in computational ethology and the physics of behaviour 10,26,27 . Maintaining sufficient resolution for pose estimation was therefore the first design constraint we required.…”

Section: Introductionmentioning

confidence: 99%

Megapixel camera arrays for high-resolution animal tracking in multiwell plates

Barlow

Feriani

Minga

et al. 2021

Preprint

View full text Add to dashboard Cite

Tracking small laboratory animals such as flies, fish, and worms is used for phenotyping in neuroscience, genetics, disease modelling, and drug discovery. Current imaging systems are limited either in spatial resolution or throughput. A system capable of imaging a large number of animals with sufficient resolution to estimate their pose would enable a new class of experiments where detailed behavioural differences are quantified but at a scale where hundreds of treatments can be tested simultaneously. Here we report a new imaging system consisting of an array of six 12-megapixel cameras that can simultaneously record from all the wells of a 96-well plate with a resolution of 80 pixels/mm at 25 frames per second. We show that this resolution is sufficient to estimate the pose of nematode worms including head identification and to extract high-dimensional phenotypic fingerprints. We illustrate the potential application of the system in three domains with a study of behavioural variability across wild isolates, experiments on the escape response and sensitisation of worms to repeated blue light stimulation, and the utility of this stimulus for phenotyping disease models. Because the system is compatible with standard multiwell plates, it makes computational ethological approaches accessible in existing high-throughput pipelines and greatly increases the scale of possible phenotypic screening experiments in C. elegans.

show abstract

WormPose: Image synthesis and convolutional networks for pose estimation inC. elegans

Cited by 10 publications

References 28 publications

Maximally predictive ensemble dynamics from data

Maximally predictive ensemble dynamics from data

Maximally predictive ensemble dynamics from data

Megapixel camera arrays for high-resolution animal tracking in multiwell plates

Contact Info

Product

Resources

About