Identifying geographically differentiated features of Ethopian Nile tilapia (Oreochromis niloticus) morphology with machine learning.

Esayas Alemayehu,Manuel Curto,Papius Tibihika,Khanh N.Q. Le,Harald Meimberg,Paul Meulenbroek,Wilfried Wöber,Peter Sykacek,Lars Mehnen

doi:10.1371/journal.pone.0249593

Abstract

Visual characteristics are among the most important features for characterizing the phenotype of biological organisms. Color and geometric properties define population phenotype and allow assessing diversity and adaptation to environmental conditions. To analyze geometric properties classical morphometrics relies on biologically relevant landmarks which are manually assigned to digital images. Assigning landmarks is tedious and error prone. Predefined landmarks may in addition miss out on information which is not obvious to the human eye. The machine learning (ML) community has recently proposed new data analysis methods which by uncovering subtle features in images obtain excellent predictive accuracy. Scientific credibility demands however that results are interpretable and hence to mitigate the black-box nature of ML methods. To overcome the black-box nature of ML we apply complementary methods and investigate internal representations with saliency maps to reliably identify location specific characteristics in images of Nile tilapia populations. Analyzing fish images which were sampled from six Ethiopian lakes reveals that deep learning improves on a conventional morphometric analysis in predictive performance. A critical assessment of established saliency maps with a novel significance test reveals however that the improvement is aided by artifacts which have no biological interpretation. More interpretable results are obtained by a Bayesian approach which allows us to identify genuine Nile tilapia body features which differ in dependence of the animals habitat. We find that automatically inferred Nile tilapia body features corroborate and expand the results of a landmark based analysis that the anterior dorsum, the fish belly, the posterior dorsal region and the caudal fin show signs of adaptation to the fish habitat. We may thus conclude that Nile tilapia show habitat specific morphotypes and that a ML analysis allows inferring novel biological knowledge in a reproducible manner.

Highlights

Visual analysis and use of anatomic features has a long tradition in biology [1,2,3]
To rule out that non biological information in the images about lake impedes drawing reliable biological conclusions, we provide a careful assessment of image features which are extracted by Gaussian process latent variable models (GP-LVM) and convolutional neuronal networks (CNNs)
The automatic relevance determination (ARD) metrics of Gaussian process classifiers (GPC) and HMC-MLP will be used to identify generalized Procrustes analysis (GPA) transformed landmarks and Gaussian process (GP)-LVM image regions of Nile tilapia which our analysis considers relevant for predicting sample origin

Summary

Introduction

Visual analysis and use of anatomic features has a long tradition in biology [1,2,3]. Visual features which are distinct for groups of specimen images provide important information for biological systematics, paleontology, evolutionary, developmental and conservation biology. Classical morphology uses landmarks to define phenotype and modern morphometrics [4] to remove confounding factors that could otherwise impede drawing robust conclusions. Fish morphology allows to discriminate genera, species, populations, and even individuals [11]. Morphology allows studying the response of shape to environmental and ecological factors such as trophic behavior [12]. Morphology can quantify relationships between different species [13, 14] and adaptation of body shape to environmental change [15, 16]. Fish morphology is known to be an expression of ecological interactions [17]

Methods

Results

Conclusion