dictyNews

Electronic Edition

Volume 45, number 31

December 6, 2019



Please submit abstracts of your papers as soon as they have been

accepted for publication by sending them to [log in to unmask]

or by using the form at

http://dictybase.org/db/cgi-bin/dictyBase/abstract_submit.



Back issues of dictyNews, the Dicty Reference database and other

useful information is available at dictyBase - http://dictybase.org.



Follow dictyBase on twitter:

http://twitter.com/dictybase





=========

Abstracts

=========





Generative Adversarial Networks for Augmenting Training Data of 

Microscopic Cell Images



Piotr Baniukiewicz, E. Josiah Lutton, Sharon Collier and Till Bretschneider



Department of Computer Science, Zeeman Institute, University of Warwick, 

Coventry, United Kingdom





Front. Comput. Sci., 26 November 2019 

https://doi.org/10.3389/fcomp.2019.00010



Generative adversarial networks (GANs) have recently been successfully 

used to create realistic synthetic microscopy cell images in 2D and predict 

intermediate cell stages. In the current paper we highlight that GANs can 

not only be used for creating synthetic cell images optimized for different 

fluorescent molecular labels, but that by using GANs for augmentation of 

training data involving scaling or other transformations the inherent length 

scale of biological structures is retained. In addition, GANs make it possible 

to create synthetic cells with specific shape features, which can be used, for 

example, to validate different methods for feature extraction. Here, we apply 

GANs to create 2D distributions of fluorescent markers for F-actin in the cell 

cortex of Dictyostelium cells (ABD), a membrane receptor (cAR1), and a 

cortex-membrane linker protein (TalA). The recent more widespread use of 

3D lightsheet microscopy, where obtaining sufficient training data is 

considerably more difficult than in 2D, creates significant demand for novel 

approaches to data augmentation. We show that it is possible to directly 

generate synthetic 3D cell images using GANs, but limitations are excessive 

training times, dependence on high-quality segmentations of 3D images, and 

that the number of z-slices cannot be freely adjusted without retraining the 

network. We demonstrate that in the case of molecular labels that are highly 

correlated with cell shape, like F-actin in our example, 2D GANs can be used 

efficiently to create pseudo-3D synthetic cell data from individually generated 

2D slices. Because high quality segmented 2D cell data are more readily 

available, this is an attractive alternative to using less efficient 3D networks.





submitted by:  Till Bretschneider  [[log in to unmask]]

——————————————————————————————————————





Basic-hydrophobic (BH) sites are localized in conserved positions inside and 

outside of PH domains and affect localization of Dictyostelium myosin 1s



Hanna Brzeska, Jesus Gonzalez, Edward D. Korn and Margaret A. Titus 





Molecular Biology of the Cell, in press



Myosin 1s have critical roles in linking membranes to the actin cytoskeleton 

via direct binding to acidic lipids. Lipid binding may occur through PIP3/PIP2-

specific PH domains or non-specific ionic interactions involving basic-

hydrophobic BH sites but the mechanism of myosin 1s distinctive lipid targeting 

is poorly understood.  Now we show that PH domains occur in all Dictyostelium 

myosin 1s and that the BH sites of Myo1A, B, C, D and F are in conserved 

positions near the b3/b4 loops of their PH domains. In spite of these shared 

lipid binding sites, we observe significant differences in myosin 1s highly 

dynamic localizations. All myosin 1s except Myo1A are present in 

macropinocytic structures but only Myo1B and Myo1C are enriched at the 

edges of macropinocytic cups and associate with the actin in actin waves.  In 

contrast, Myo1D, E and F are enclosed by the actin wave.  Mutations of BH 

sites affect localization of all Dictyostelium myosin 1s. Notably, mutation of the 

BH site located within the PH domains of PIP3- specific Myo1D and Myo1F 

completely eradicates membrane binding.  Thus, BH sites are important 

determinants of motor targeting and may have a similar role in the localization 

of other myosin 1s.





Submitted by: Hanna Brzeska [[log in to unmask]]

==============================================================

[End dictyNews, volume 45, number 31]