A computational methodology to study cells morphology in developing embryos
Abstract
Introduction. The understanding of how embryos develop seeks to answer fundamental biological questions such as cell and tissue interaction during morphogenesis. Recently, new optical techniques allow the observation of embryonic development as a whole. However, the methodologies to make observe and quantify in full embryos are still a challenge. In this work, we propose a methodology to observe full embryo using a 2-lens lightsheet microscopy, performing automatic tracking and describing population morphology. Material and Methods. Annual fish embryos (~ 1.1 mm) were microinjected with EGFP targeting plasma membrane at one cell stage. During epiboly, a representative staining embryo was visualized using 2-lens lightsheet microscopy. Stacks of 0.65x0.65x3 μm images were obtained every 15 min for 2 days. Beads-based registration and fusion was performed, and cells were segmented by an intensity threshold on laplacian filtered stacks (σ=5). Detected cells were then individually tracked using a Linear Assignment Problem (LAP) tracking algorithm. Results. From an estimated global population of 150 cells populations we are able to track 82 (55%) without long gaps. To further study this population we describe cell shape by using elongation (ratio between main and secondary axis) and measure a time series distance (time warp metric) identifying two clusters. We are currently relating these cluster with spatial cell locations. Overall, we present a complete methodology going from microscopy to spatial-temporal patterns identification on live embryos. Funding: ICM P09-015-F, FONDECYT (11170761, 1161274, 1151029, 3130598, 3140447, 11161033, 3140444), CONICYT PAI 79170126, CONICYT PIA ACT1402, FONDAP 15150012, FONDEQUIP (EQM130051, EQM140119).
Más información
Fecha de publicación: | 2019 |
Año de Inicio/Término: | 3-04-2019 |