Biology subjects

Evans, B. D.

Publications and source records attributed to Evans, B. D..

6 recordsLinked to original sources

bioRxiv2026

Factorization and spatial encodings: a hypothesis about the foundations of the genomic code

The genomic mechanisms that efficiently encode the initial architecture and synaptic connectivity of neural circuits remain poorly understood. We hypothesise that two primary mechanisms -- spatial encoding and factorisation -- enable a limited genome to initialise networks of billions of neurons. Spatial encoding, a form of indirect representation, compresses neural network parameters while enforcing structural continuity. Complementarily, we introduce a factorisation mechanism inspired by reaction-diffusion models, comprising a spatially invariant reaction rule and spatially variant diffusion dynamics. Based on this insight, we can efficiently abstract a neural network layer as a spatially invariant weight kernel (or filter) and its spatially variant transformations along the spatial dimensions. Thus, coupling this variant-invariant decomposition with spatial encodings can substantially reduce the size of the solution space explored by the genome. In addition, we show that this coupling leads to efficient initialisation of cortical maps, such as V1 orientation maps, and neural networks with the ability to generalise. In summary, coupling factorisation and spatial encodings can offer functional advantages to the evolving genome.

neuroscience↗

bioRxiv2022

Survivor bias drives overestimation of stability in reconstructed ancestral proteins

Ancestral sequence reconstruction has been broadly employed over the past two decades to probe the evolutionary history of life. Many ancestral sequences are thermostable, supporting the "hot-start" hypothesis for lifes origin. Recent studies have observed thermostable ancient proteins that evolved in moderate temperatures. These effects were ascribed to "consensus bias". Here, we propose that "survivor bias" provides a complementary rationalisation for ancestral protein stability in alignment-based methods. As thermodynamically unstable proteins will be selected against, ancestral or consensus sequences derived from extant sequences are selected from a dataset biased towards the more stabilising amino acids in each position. We thoroughly explore the presence of survivor bias using a highly parameterizable in silico model of protein evolution that tracks stability at the population, protein, and amino acid levels. We show that ancestors and consensus sequences derived from populations evolved under selective pressure for stability throughout their history are significantly biased toward thermostability. Our work proposes a complementary explanation of the origin of thermostability in the burgeoning engineering tools of ancestral sequence reconstruction and consensuses. It provides guidance for the thorough derivation of conclusions from future ancestral sequence reconstruction work.

evolutionary biology↗

bioRxiv2022

The role of capacity constraints in Convolutional Neural Networks for learning random versus natural data

Convolutional neural networks (CNNs) are often described as promising models of human vision, yet they show many differences from human abilities. We focus on a superhuman capacity of top-performing CNNs, namely, their ability to learn very large datasets of random patterns. We verify that human learning on such tasks is extremely limited, even with few stimuli. We argue that the performance difference is due to CNNs overcapacity and introduce biologically inspired mechanisms to constrain it, while retaining the good test set generalisation to structured images as characteristic of CNNs. We investigate the efficacy of adding noise to hidden units activations, restricting early convolutional layers with a bottleneck, and using a bounded activation function. Internal noise was the most potent intervention and the only one which, by itself, could reduce random data performance in the tested models to chance levels. We also investigated whether networks with biologically inspired capacity constraints show improved generalisation to out-of-distribution stimuli, however little benefit was observed. Our results suggest that constraining networks with biologically motivated mechanisms paves the way for closer correspondence between network and human performance, but the few manipulations we have tested are only a small step towards that goal.

neuroscience↗

bioRxiv2022

Bespoke data augmentation and network construction enable image classification on small microscopy datasets

Recent work has indicated a need for increased temporal resolution for studies of the early chick brain. Over a 10-hour period, the developmental potential of progenitor cells in the HH10 brain changes, and concomitantly, the brain undergoes subtle changes in morphology. We asked if we could train a deep convolutional neural network to sub-stage HH10 brains from a small dataset (<200 images). By augmenting our images with a combination of biologically informed transformations and data-driven preprocessing steps, we successfully trained a classifier to sub-stage HH10 brains to 87.1% test accuracy. To determine whether our classifier could be generally applied, we re-trained it using images (<250) of randomised control and experimental chick wings, and obtained similarly high test accuracy (86.1%). Saliency analyses revealed that biologically relevant features are used for classification. Our strategy enables training of image classifiers for various applications in developmental biology with limited microscopy data. SUMMARY STATEMENTWe train a deep convolutional network that can be generally applied to accurately classify chick embryos from images. Saliency analyses show that classification is based on biologically relevant features.

developmental biology↗

bioRxiv2021

Biological convolutions improve DNN robustness to noise and generalisation

Deep Convolutional Neural Networks (DNNs) have achieved superhuman accuracy on standard image classification benchmarks. Their success has reignited significant interest in their use as models of the primate visual system, bolstered by claims of their architectural and representational similarities. However, closer scrutiny of these models suggests that they rely on various forms of shortcut learning to achieve their impressive performance, such as using texture rather than shape information. Such superficial solutions to image recognition have been shown to make DNNs brittle in the face of more challenging tests such as noise-perturbed or out-of-domain images, casting doubt on their similarity to their biological counterparts. In the present work, we demonstrate that adding fixed biological filter banks, in particular banks of Gabor filters, helps to constrain the networks to avoid reliance on shortcuts, making them develop more structured internal representations and more tolerant to noise. Importantly, they also gained around 20 35% improved accuracy when generalising to our novel out-of-domain test image sets over standard end-to-end trained architectures. We take these findings to suggest that these properties of the primate visual system should be incorporated into DNNs to make them more able to cope with real-world vision and better capture some of the more impressive aspects of human visual perception such as generalisation.

neuroscience↗

bioRxiv2020

Vangl2 regulates the dynamics of Wnt cytonemes in vertebrates

The Wnt signalling network regulates cell proliferation and cell differentiation as well as migration and polarity in development of multicellular organisms. However, it is still unclear how distribution of Wnt ligands is precisely controlled to fulfil all of these different functions. Here, we show that the four-pass transmembrane protein Vangl2 occupies a central role in determining the distribution of Wnt by cytonemes in vertebrate tissue. In zebrafish epiblast cells, mouse intestinal telocytes and human gastric cancer cells, activation of Vangl2 leads to the generation of fewer but extremely long cytonemes, which start to branch and deliver Wnt protein to multiple cells. The Vangl2-activated cytonemes increase Wnt/{beta}-catenin signalling in the surrounding cells. Concordantly, inhibition of Vangl2 function leads to the formation of shorter cytonemes and reduced paracrine Wnt/{beta}-catenin signal activation. A mathematical model simulating the observed Vangl2 functions on cytonemes in zebrafish gastrulation predicts an anterior shift of the morphogenetic signalling gradient, altered tissue patterning, and a loss of the sharpness of tissue domains. We confirmed these predictions during anteroposterior patterning in the zebrafish neural plate. In summary, we show that Vangl2 - a core member of the PCP signalling component - is fundamental to paracrine Wnt/{beta}-catenin signalling by controlling cytoneme behaviour in vertebrate development and tissue homeostasis.

developmental biology↗