Biology subjects

Whittaker, J.

Publications and source records attributed to Whittaker, J..

7 recordsLinked to original sources

bioRxiv2026

Genomic foundation models do not yet boost common variant discovery beyond standard functional annotations

Genomic foundation models (GFMs) use large-scale AI to model genetic sequences and their influence on molecular phenotypes. While GFMs have shown promise in rare variant analysis, their utility for common variant discovery remains untested. Using the surrogate functional false discovery rate framework for functionally informed GWAS discovery, we compared standard functional annotations (baselineLD) with annotations derived from three leading GFMs: the sequence-to-function models AlphaGenome and Borzoi, and the genomic language model GPN-Star. Across 27 UK Biobank traits, GPN-Star decreased locus discovery relative to a standard GWAS, whereas AlphaGenome and Borzoi increased discovery but by less than baselineLD (medians 1.8%, 3.8%, and 5.1%). When we combined the GFM and baselineLD annotations, there was no increase over baselineLD alone (median changes -0.1% to 0.2%), suggesting that AlphaGenome and Borzoi mostly capture information already in baselineLD. Thus, GFMs may require larger corpora or trait-specific feature engineering to improve discovery beyond standard functional annotations.

genomics↗

bioRxiv2026

A family of metabolite damage-control phosphatases modulates ROS-induced autophagy and plant stress resilience

Autophagy is an intracellular recycling pathway with profound impacts on development, growth, and stress tolerance in eukaryotes. Therefore, unravelling signalling mechanisms that modulate the process has broad applicability. Here, we characterised a small organic molecule as an enhancer of autophagy across diverse plant lineages. Through an in vivo photoaffinity proteomics approach in Arabidopsis thaliana seedlings, we identified Domain of Unknown Function 89 (DUF89) proteins as molecular targets. Hereafter, we refer to these proteins as Reactive Metabolite Damage-Control Phosphatases (RMDPs), which are metabolite-repair phosphatases that clear reactive products. The new autophagy modulator is called RMDP inhbitor-1 (RMDPi-1) due to its demonstrated inhibitory effect on AtRMDP1 activity in vitro. Binding was further confirmed through co-incubation that thermostabilised the protein. Crystal structures of AtRMDP1 were obtained via X-ray diffraction, confirming binding of RMDPi-1 in the catalytic site and revealing a flexible region near the binding pocket that was absent in previous models. Molecular docking predicted two possible binding modes consistent with the observed electron density. RMDPi-1 treatment and the phosphatase gene knockout increased autophagic flux in Arabidopsis and Chlamydomonas reinhardtii, establishing a conserved effect. The observed autophagy response is linked to a spike in reactive oxygen species (ROS). AtRMDP knockout also results in higher ROS tolerance and greater biomass in Arabidopsis seedlings. Targeted and untargeted metabolomic experiments revealed that both pharmacological and genetic suppression of RMDP leads to accumulation of glycating agents and phosphate sugars, associating loss of RMDP activity to the ROS stress that induces alternate ROS-scavenging mechanisms, and autophagy. Taken together, our integrated chemical genetics approach reveals enhanced autophagy via inhibition of RMDP family members, which impacts on plant stress tolerance via conserved metabolite damage-control functions.

plant biology↗

bioRxiv2026

A DNA foundation model predicts osteoporosis risk genes without proximity bias

Targets supported by human genetic associations are more than twice as likely to progress from clinical development to approval. Genome-wide association studies are the largest source of genetic evidence for disease risk but linking non-coding variants to effector genes remains a significant barrier to identifying causal targets. Current gene-mapping approaches suffer from proximity bias, largely ignoring distal genes. Here we introduce Rosalind, a DNA foundation model fine-tuned on human genetic variation from GTEx, that directly predicts variant-gene regulatory relationships from sequence without relying on nearest-gene heuristics. We demonstrate Rosalinds accuracy through extensive benchmarking, apply it to multiple complex traits to establish broad utility, and provide experimental validation in osteoporosis using a translational osteoblast assay. We demonstrate that genes distal to osteoporosis risk variants were significantly more likely to alter a bone formation phenotype than nearest genes. Together, these results highlight deep learning-based regulatory models as a general and scalable framework for translating novel genetic insights to drug discovery.

genetics↗

bioRxiv2025

Systematic comparison of colocalization methods using protein quantitative trait loci

Colocalization is frequently performed as a step to triage findings from genetic investigations linking molecular and disease data. However, the reliability and consistency of the various colocalization methods is not well understood. In particular, it is unclear whether a non-colocalization result should be taken as a definitive sign that two traits do not share a genetic cause in the region of interest, or merely as suggestive evidence. We use protein QTL data to benchmark four colocalization methods (coloc, coloc-SuSiE, prop-coloc, and colocPropTest), considering whether the methods conclude there is colocalization between datasets representing associations with the same protein. We consider a baseline scenario in which the associations come from the same dataset split in half at random, and scenarios in which the associations are with the same protein, but measured on different platforms or in different populations. In the baseline scenario, all methods report colocalization for the majority of proteins. In other scenarios, methods do not consistently report colocalization, they often report non-colocalization, and they often disagree. In the worst-case scenario, colocalization was only agreed by all four methods for 20% of proteins, despite our experiment being constructed to select for cases where colocalization is expected. This suggests caution in performing and interpreting colocalization analyses is warranted.

genetics↗

bioRxiv2022

Stabilization of a Protein by a Single Halogen-Based Aromatic Amplifier

The utility of halogenation in protein design is investigated by a combination of quantitative atomistic simulations and experiment. The approach is applied to insulin, a small, therapeutically relevant domain amenable to simulation and semi-synthesis. In a singly halogenated aromatic ring, the simulations predicted regiospecific inductive effects to modulate multiple surrounding electrostatic (weakly polar) interactions, thereby amplifying changes in thermodynamic stability. In accordance with the simulations, stabilization of insulin is demonstrated by single halogen atoms at the ortho position of an invariant phenylalanine (2-F-PheB24, 2-Cl-PheB24 and 2-Br-PheB24; {Delta}{Delta}Gu = -0.5 to -1.0 kcal/mol) located at the edge of a protein crevice. Corresponding meta and para substitutions have negligible effects. The ortho-modified insulin analogs exhibit enhanced resistance to fibrillation above room temperature and retain biological activity in mammalian cells and in a rat model of diabetes mellitus. Consequently, halogen-based stabilization of insulin and other therapeutic proteins may provide a biophysical strategy to circumvent the requirement for a distribution "cold chain" in the developing world and enhance the shelf life of pharmaceutical formulations.

biochemistry↗

bioRxiv2020

Multi-proxy evidence for the impact of the Storegga Slide Tsunami on the early Holocene landscapes of the southern North Sea

Doggerland was a land mass occupying an area currently covered by the North Sea until marine inundation took place during the mid-Holocene, ultimately separating the British land mass from the rest of Europe. The Storegga Slide, which triggered a tsunami reflected in sediment deposits in the Northern North Sea, North East coastlines of the British Isles and across the North Atlantic, was a major event during this transgressive phase. The spatial extent of the Storegga tsunami however remains unconfirmed because to date no direct evidence for the event has been recovered from the southern North Sea. We present evidence that Storegga associated deposits occur in the southern North Sea. Palaeo-river systems have been identified using seismic survey in the southwestern North Sea and sedimentary cores extracted to track the Mid Holocene inundation. At the head of one palaeo-river system near the Outer Dowsing Deep, the Southern River, we observed an abrupt and catastrophic inundation stratum. Based on lithostratigraphic, macro and microfossils and sedimentary ancient DNA (sedaDNA) evidence, supported by optical stimulation luminescence (OSL) and radiocarbon dating, we conclude these deposits were a result of the Storegga event. Seismic identification of this stratum to adjacent cores indicated diminished traces of the tsunami, largely removed by subsequent erosional processes. Our results demonstrate the catastrophic impact of Storegga within this area of the Southern North Sea, but indicate that these effects were temporary and likely localized and mitigated by the dense woodland and topography of the area. We conclude clear physical remnants of the wave are likely to be restricted to inland basins and incised river valley systems.

paleontology↗

bioRxiv2017

Causal Analyses, Statistical Efficiency And Phenotypic Precision Through Recall-By-Genotype Study Design

Genome-wide association studies have been useful in identifying common genetic variants related to a variety of complex traits and diseases; however, they are often limited in their ability to inform about underlying biology. Whilst bioinformatics analyses, studies of cells, animal models and applied genetic epidemiology have provided some understanding of genetic associations or causal pathways, there is a need for new genetic studies that elucidate causal relationships and mechanisms in a cost-effective, precise and statistically efficient fashion. We discuss the motivation for and the characteristics of the Recall-by-Genotype (RbG) study design, an approach that enables genotype-directed deep-phenotyping and improvement in drawing causal inferences. Specifically, we present RbG designs using single and multiple variants and discuss the inferential properties, analytical approaches and applications of both. We consider the efficiency of the RbG approach, the likely value of RbG studies for the causal investigation of disease aetiology and the practicalities of incorporating genotypic data into population studies in the context of the RbG study design. Finally, we provide a catalogue of the UK-based resources for such studies, an online tool to aid the design of new RbG studies and discuss future developments of this approach.

genetics↗