Powerful and interpretable behavioural features for quantitative phenotyping ofCaenorhabditis elegans

Abstract

Behaviour is a sensitive and integrative readout of nervous system function and therefore an attractive measure for assessing the effects of mutation or drug treatment on animals. Video data provide a rich but high-dimensional representation of behaviour, and so the first step of analysis is often some form of tracking and feature extraction to reduce dimensionality while maintaining relevant information. Modern machine-learning methods are powerful but notoriously difficult to interpret, while handcrafted features are interpretable but do not always perform as well. Here, we report a new set of handcrafted features to compactly quantify Caenorhabditis elegans behaviour. The features are designed to be interpretable but to capture as much of the phenotypic differences between worms as possible. We show that the full feature set is more powerful than a previously defined feature set in classifying mutant strains. We then use a combination of automated and manual feature selection to define a core set of interpretable features that still provides sufficient power to detect behavioural differences between mutant strains and the wild-type. Finally, we apply the new features to detect time-resolved behavioural differences in a series of optogenetic experiments targeting different neural subsets.This article is part of a discussion meeting issue ‘Connectome to behaviour: modelling C. elegans at cellular resolution’.

Systemic muscle wasting and coordinated tumour response drive tumourigenesis

“…When the head was correctly identified it was stable over the course of the video. 256 defined features, previously determined to be useful in classifying behaviour in C.elegans , were extracted from the tracking data and compared between conditions 13 . Correction for multiple testing was applied using the Benjamini–Hochberg procedure to control the false discovery rate to 0.05 31 .…”

“…1f-j). Functional locomotion defects—as measured by video tracking analysis 12,13 —were observed in ras1 G12V ; csk -/- animals raised on a HSD (Fig. 1i).…”

A terminal selector prevents a Hox transcriptional switch to safeguard motor neuron identity throughout life

“…We predicted that if ectopic expression of VD and VC genes contributes to locomotion defects, then unc-3 (n3435) mutants would display more severe locomotion defects than unc-3 (n3435); lin-39 (n1760) double mutants. To test this, we performed high-resolution behavioral analysis of freely moving adult (day 1) C. elegans animals using automated multi-worm tracking technology [52,53]. This analysis can quantitate multiple features related to C. elegans locomotion (e.g., speed, crawling amplitude, curvature, pause, forward and backward locomotion) and, most importantly, each feature can be localized to a specific part of the nematode's body (e. g., head, mid-body, tail).…”

“…All videos were analyzed using Tierpsy Tracker [73] to extract each worm's position and posture over time. These postural data were then converted into a set of behavioral features as previously described [53]. From the total set of features, we only considered 48 that are related to midbody posture and motion, as well as the midbody width (see Table S3 for feature descriptions and their average values for each strain).…”

Establishment and maintenance of motor neuron identity via temporal modularity in terminal selector function

“…As controls, we used animals carrying: (a) the endogenous cfi-1 reporter allele (mNG::AID::cfi-1), (b) a putative null cfi-1 allele (ot786) [42], in which cfi-1 activity is affected in MNs and other neuron types of the motor circuit [32,45], and (c) a deletion of the distal enhancer (mNG::AID::cfi-1 Δenhancer(769 bp) ), in which both initiation and maintenance of cfi-1 are abrogated in nerve cord MNs (Figure 6A). We performed a high-resolution behavioral analysis of freely moving adult (day 2) animals of the above genotypes using automated multi-worm tracking technology [54,55]. We found several features related to C. elegans locomotion (e.g., body curvature, velocity) severely affected in cfi-1 (ot786) putative null animals (Figure 6B-G).…”