Competitive fitness analysis using Convolutional Neural Network

Abstract

We developed a procedure for estimating competitive fitness by using Caenorhabditis elegans as a model organism and a Convolutional Neural Network (CNN) as a tool. Competitive fitness is usually the most informative fitness measure, and competitive fitness assays often rely on green fluorescent protein (GFP) marker strains. CNNs are a class of deep learning neural networks, which are well suited for image analysis and object classification. Our model analyses involved image classification of nematodes as wild-type vs. GFP-expressing, and counted both categories. The performance was analyzed with (i) precision and recall parameters, and (ii) comparison of the wild-type frequency calculated from the model against that obtained by visual scoring of the same images. The average precision and recall varied from 0.79 to 0.87 and from 0.84 to 0.92, respectively, depending on worm density in the images. Compared with manual counting, the model decreased counting time at least 20-fold while preventing human errors. Given the rapid development in the field of CNN, the model, which is fully available on GitHub, can be further optimized and adapted for other image-based uses.

Reproductive system, temperature, and genetic background effects in experimentally evolving populations of Caenorhabditis elegans

“…The sample was mounted on a glass slide and covered with a cover slip. Approximately 10 non-overlapping pictures were done to estimate the initial proportion of focal vs GFP populations and analysed using the automated method based on machine learning taught to distinguish GFP from non-GFP animals [ 35 ].…”

“…The filtered liquid (with larvae only) was placed in Eppendorf tubes, from which, after sedimentation, a 5 μl drop was taken and placed on a glass slide with a cover slip ( Fig 3 ). Again, 10 non-overlapping pictures were taken and were analysed using the automated method [ 35 ]. The proportion of focal (non-fluorescent) larvae in each sample was calculated.…”

Evolution of Reproductive Efficiency in Caenorhabditis elegans Under Obligatory Outcrossing

“…Ten non-overlapping pictures per slide were then taken under magnification ×40 in the Nikon Eclipse 80i microscope with BV-1A filter combination (435/10 nm excitation filter, 470 nm barrier filter, dichromatic mirror value 455 nm), equipped with Nikon Digital Sight DS-U3 camera connected to a computer with NIS-Elements software. Pictures have been analysed using a software based on machine learning, which was previously taught to recognize non-GFP from GFP animals, based on a set of pictures with manually assigned individuals (Palka et al, 2020). The algorithm analyses pictures containing both types of animals by finding nematodes and recognizing them as fluorescent or nonfluorescent.…”

“…Fig.3Nematode image from fitness assay processed by algorithm classifying them as fluorescent (marked as yellow) or non-fluorescent (markred as green). For details seePalka et al (2020) …”