Biology subjects

Gardiner, L.-J.

Publications and source records attributed to Gardiner, L.-J..

14 recordsLinked to original sources

bioRxiv2026

Macrophage signature-based prediction of cancer treatment response using MIL-attention

Predicting immunotherapy response from single-cell data remains difficult due to patient-level labels, extreme class imbalance, and highly heterogeneous macrophage states. We present a Multiple Instance Learning (MIL) framework that treats each patient as a bag of macrophage embeddings derived from a single-cell RNA foundation model. The architecture incorporates an attention-based pooling mechanism with reduced model complexity, dropout-enhanced regularization and explicit attention penalties to improve stability in small-sample regimes. To address imbalanced clinical datasets, MIL outputs are optimized with a combined focal loss and supervised contrastive objective that simultaneously sharpens class boundaries and improves representation clustering. Across three cancer datasets, this approach outperforms pseudobulk aggregation, embedding baselines and standard MIL variants. Attention-weighted attribution and transcriptional regulatory analysis reveal distinct macrophage programs, interferon and antigen-presentation networks in responders versus hypoxia-linked regulatory modules in non-responders. This shows the potential of MIL to uncover predictive and mechanistically interpretable immune states.

bioinformatics↗

bioRxiv2026

RNA foundation models enable generalizable endometriosis disease classification and stable gene-level interpretation

Endometriosis is a chronic inflammatory condition with significant diagnostic delays impacting one in ten reproductive age women worldwide. While machine learning (ML) models trained on transcriptomic data show promise for disease prediction, limited generalizability across independent patient cohorts has hindered clinical translation. Foundations models (FMs) pretrained on large-scale transcriptomic data offer promise to learn transferrable, biologically meaningful representations that could support cross-cohort predictions. We assembled a 12-cohort bulk RNA-seq benchmark (334 samples) and developed a computationally efficient pipeline to test whether FMs improve endometriosis classification, an approach not previously applied to this disease. Using AutoXAI4Omics with cohort-aware validation, we compared embeddings derived from five state-of-the-art RNA FMs against TPM baselines. In cross-cohort prediction, FM embeddings significantly improved performance, achieving a weighted F1-score of 0.83 vs. 0.68 for the baseline. To allow gene-level interpretation of FM embedding models, we introduce classified-aligned integrated gradients (CA-IG), an interpretability approach aligning gene-level attributions to the downstream classifier without end-to-end finetuning. CA-IG revealed a conserved set of predictive genes from FM embeddings across cohort-validation regimes, contrasting with unstable baseline explainability, suggesting that FM embeddings prioritized transferable disease-related signal over cohort-specific effects. These genes include novel candidates that converge on biologically plausible pathways for endometriosis.

bioinformatics↗

bioRxiv2024

AutoXAI4Omics: an Automated Explainable AI tool for Omics and tabular data

Machine learning (ML) methods have the potential of detailed insights of complex biological systems and today are increasingly used to analyse omics data for tasks such as the discovery of novel biomarkers and phenotype prediction. It can be extremely beneficial and powerful for scientists, domain experts, to easily run sophisticated, robust, and interpretable ML pipelines without the need for an in depth understanding of the code needed to train, tune, optimise ML algorithms. They can then focus on the biological interpretation and validation of the results and insights generated by ML models. Here, we present an entirely automated open-source explainable AI tool, AutoXAI4Omics, that performs classification and regression tasks from omics and tabular numerical data. AutoXAI4Omics accelerates scientific discovery by automating processes and decisions made by AI experts, e.g., selection of the best feature set, hyper-tuning of different ML algorithms and selection of the best ML model for a specific task and dataset. Prior to ML analysis AutoXAI4Omics incorporates feature filtering options that are tailored to specific omic data types. Moreover, the insights into the predictions that are provided by the tool through explainability analysis highlight associations between omic feature values and the targets under investigation e.g., predicted phenotypes, facilitating the discovery of actionable insights. AutoXAI4Omics is at: https://github.com/IBM/AutoXAI4Omics. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=188 HEIGHT=200 SRC="FIGDIR/small/586460v1_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@366327org.highwire.dtl.DTLVardef@a7b559org.highwire.dtl.DTLVardef@7319aforg.highwire.dtl.DTLVardef@9b5030_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2021

Interpreting machine learning models to investigate circadian regulation and facilitate exploration of clock function

The circadian clock is an important adaptation to life on earth. Here, we use machine learning to predict complex temporal circadian gene expression patterns in Arabidopsis. Most significantly, we classify circadian genes using DNA sequence features generated from public genomic resources, with no experimental work or prior knowledge needed. We use model explanation to rank DNA sequence features, observing transcript-specific combinations of potential circadian regulatory elements that discriminate temporal phase of expression. Model interpretation/explanation provides the backbone of our methodological advances, giving insight into biological processes and experimental design. Next, we use model interpretation to optimize sampling strategies when we predict circadian transcripts using reduced numbers of transcriptomic timepoints, saving both time and money. Finally, we predict the circadian time from a single transcriptomic timepoint, deriving novel marker transcripts that are most impactful for accurate prediction, this could facilitate the identification of altered clock function from existing datasets.

genomics↗

bioRxiv2020

Explainable AI reveals key changes in skin microbiome associated with menopause, smoking, aging and skin hydration

Alterations in the human microbiome have been observed in a variety of conditions such has asthma, gingivitis, dermatitis and cancer, and much remains to be learned about the links between the microbiome and human health. The fusion of artificial intelligence with rich microbiome datasets can offer an improved understanding of the microbiomes role in our health. To gain actionable insights it is essential to consider both the predictive power and the transparency of the models by providing explanations for the predictions. We combine the effort of collecting a corpus of leg skin microbiome samples of two healthy cohorts of women with the development of an explainable artificial intelligence (EAI) approach that provides accurate predictions of phenotypes and explanations. The explanations are expressed in terms of variations in the abundance of key microbes that drive the predictions. We predict skin hydration, subjects age, pre/post-menopausal status and smoking status from the leg skin microbiome. The key changes in microbial composition linked to skin hydration can accelerate the development of personalised treatments for healthy skin, while those associated with age may offer insights into the skin aging process. The leg microbiome signatures associated with smoking and menopausal status are consistent with previous findings from oral/respiratory tract microbiomes and vaginal microbiomes respectively. This suggests that easily accessible microbiome samples could be used to investigate health-related phenotypes, offering potential for non-invasive diagnosis and condition monitoring. Our EAI approach sets the stage for new work focused on understanding the complex relationships between microbial communities and phenotypes. Our approach can be applied to predict any conditions from microbiome samples and has the potential to accelerate the development of microbiome-based personalised therapeutics and non-invasive diagnostics.

bioinformatics↗

bioRxiv2020

Uncovering candidate genes involved in photosynthetic capacity using unexplored genetic variation in Spring Wheat

To feed an ever-increasing population we must leverage advances in genomics and phenotyping to harness the variation in wheat breeding populations for traits like photosynthetic capacity which remains unoptimized. Here we survey a diverse set of wheat germplasm containing elite, introgression and synthetic derivative lines uncovering previously uncharacterised variation. We demonstrate how strategic integration of exotic material alleviates the D genome genetic bottleneck in wheat, increasing SNP rate by 62% largely due to Ae. tauschii synthetic wheat donors. Across the panel, 67% of the Ae. tauschii donor genome is represented as introgressions in elite backgrounds. We show how observed genetic variation together with hyperspectral reflectance data can be used to identify candidate genes for traits relating to photosynthetic capacity using association analysis. This demonstrates the value of genomic methods in uncovering hidden variation in wheat and how that variation can assist breeding efforts and increase our understanding of complex traits.

plant biology↗

bioRxiv2019

A framework for gene mapping in wheat demonstrated using the Yr7 yellow rust resistance gene

We used three approaches to map the yellow rust resistance gene Yr7 and identify associated SNPs in wheat. First, we used a traditional QTL mapping approach using a double haploid (DH) population and mapped Yr7 to a low-recombination region of chromosome 2B. To fine map the QTL, we then used an association mapping panel. Both populations were SNP array genotyped allowing alignment of QTL and genome-wide association scans based on common segregating SNPs. Analysis of the association panel spanning the QTL interval, narrowed the interval down to a single haplotype block. Finally, we used mapping-by-sequencing of resistant and susceptible DH bulks to identify a candidate gene in the interval showing high homology to a previously suggested Yr7 candidate and to populate the Yr7 interval with a higher density of polymorphisms. We highlight the power of combining mapping-by-sequencing, delivering a complete list of gene-based segregating polymorphisms in the interval with the high recombination, low LD precision of the association mapping panel. Our mapping-by-sequencing methodology is applicable to any trait and our results validate the approach in wheat, where with a near complete reference genome sequence, we are able to define a small interval containing the causative gene.\n\nHighlightWe show progression from genetic mapping to mapping-by-sequencing and the overlap of defined intervals by each approach culminating with interval refinement and identification of a candidate gene for disease resistance.

genetics↗

bioRxiv2019

Identification of nitrogen-dependent QTL and underlying genes for root system architecture in hexaploid wheat

The root system architecture (RSA) of a crop has a profound effect on the uptake of nutrients and consequently the potential yield. However, little is known about the genetic basis of RSA and resource adaptive responses in wheat (Triticum aestivum L.). Here, a high-throughput germination paper plant phenotyping system was used to identify seedling traits in a wheat doubled haploid mapping population, Savannah x Rialto. Significant genotypic and nitrate-N treatment variation was found across the population for seedling traits with distinct trait grouping for root size-related traits and root distribution-related traits. Quantitative trait locus (QTL) analysis identified a total of 59 seedling trait QTLs. Across two nitrate treatments, 27 root QTLs were specific to the nitrate treatment. Transcriptomic analyses for one of the QTLs on chromosome 2D found under low nitrate conditions was pursued revealing gene enrichment in N-related biological processes and 17 candidate up-regulated genes with possible involvement in a root angle response. Together, these findings provide genetic insight into root system architecture and plant adaptive responses to nitrate and provide targets that could help improve N capture in wheat.

plant biology↗

bioRxiv2019

Reduced chromatin accessibility underlies gene expression differences in homologous chromosome arms of hexaploid wheat and diploid Aegilops tauschii

Polyploidy has been centrally important in driving the evolution of plants, and leads to alterations in gene expression that are thought to underlie the emergence of new traits. Despite the common occurrence of these global patterns of altered gene expression in polyploids, the mechanisms involved are not well understood. Using a precise framework of highly conserved syntenic genes on hexaploid wheat chromosome 3DL and its progenitor 3L chromosome arm of diploid Aegilops tauschii, we show that 70% of these genes exhibited proportionally reduced gene expression, in which expression in the hexaploid context of the 3DL genes was approximately 40% of the levels observed in diploid Ae. tauschii. Many genes showing elevated expression during later stages of grain development in wheat compared to Ae. tauschii. Gene sequence and methylation differences accounted for only a few cases of differences in gene expression. In contrast, large scale patterns of reduced chromatin accessibility of genes in the hexaploid chromosome arm compared to its diploid progenitor were correlated with observed overall reduction in gene expression and differential gene expression. Therefore, that an overall reduction in accessible chromatin underlies the major differences in gene expression that results from polyploidization.

genomics↗

bioRxiv2019

Analysis of the recombination landscape of hexaploid bread wheat reveals genes controlling recombination and gene conversion frequency

Sequence exchange between homologous chromosomes through crossing over and gene conversion is highly conserved among eukaryotes, contributing to genome stability and genetic diversity. Lack of recombination limits breeding efforts in crops, therefore increasing recombination rates can reduce linkage-drag and generate new genetic combinations. We use computational analysis of 13 recombinant inbred mapping populations to assess crossover and gene conversion frequency in the hexaploid genome of wheat (Triticum aestivum). We observe that high frequency crossover sites are shared between populations and that closely related parental founders lead to populations with more similar crossover patterns. We demonstrate that gene conversion is more prevalent and covers more of the genome in wheat than in other plants, making it a critical process in the generation of new haplotypes, particularly in centromeric regions where crossovers are rare. We have identified QTL for altered gene conversion and crossover frequency and confirm functionality for a novel RecQ helicase gene that belongs to an ancient clade that is missing in some plant lineages including Arabidopsis. This is the first gene to be demonstrated to be involved in gene conversion in wheat. Harnessing the RecQ helicase has the potential to break linkage-drag utilizing widespread gene conversions.

genetics↗

bioRxiv2018

Elucidating the genetic basis of biomass accumulation and radiation use efficiency in spring wheat and its role in yield potential

One of the major challenges for plant scientists is increasing wheat (Triticum aestivum) yield potential (YP). A significant bottleneck for increasing YP is achieving increased biomass through optimization of Radiation Use Efficiency (RUE) along the crop cycle. Exotic material such as landraces and synthetic wheat has been incorporated into breeding programs in an attempt to alleviate this, however their contribution to YP is still unclear. To understand the genetic basis of biomass accumulation and RUE we applied genome-wide association study (GWAS) to a panel of 150 elite spring wheat genotypes including many landrace and synthetically derived lines. The panel was evaluated for 31traits over two years under optimal growing conditions and genotyped using the 35K Wheat Breeders array. Marker-trait-association identified 94 SNPs significantly associated with yield, agronomic and phenology related traits along with RUE and biomass at various growth stages that explained 7-17 % of phenotypic variation. Common SNP markers were identified for grain yield, final biomass and RUE on chromosomes 5A and 7A. Additionally we show that landrace and synthetic derivative lines showed higher thousand grain weight (TGW), biomass and RUE but lower grain number (GNO) and harvest index (HI). Our work demonstrates the use of exotic material as a valuable resource to increase YP. It also provides markers for use in marker assisted breeding to systematically increase biomass, RUE and TGW and avoid the TGW/GNO and BM/HI trade-off. Thus, achieving greater genetic gains in elite germplasm while also highlighting genomic regions and candidate genes for further study.

plant biology↗

bioRxiv2018

Integrating genomic resources to present full gene and promoter capture probe sets for bread wheat

BackgroundWhole genome shotgun re-sequencing of wheat is expensive because of its large, repetitive genome. Moreover, sequence data can fail to map uniquely to the reference genome making it difficult to unambiguously assign variation. Re-sequencing using target capture enables sequencing of large numbers of individuals at high coverage to reliably identify variants associated with important agronomic traits.\n\nResultsWe present and validate two gold standard capture probe sets for hexaploid bread wheat, a gene and a promoter capture, which are designed using recently developed genome sequence and annotation resources. The captures can be combined or used independently. We demonstrate that the capture probe sets effectively enrich the high confidence genes and promoters that were identified in the genome alongside a large proportion of the low confidence genes and promoters. Finally, we demonstrate successful sample multiplexing that allows generation of adequate sequence coverage for SNP calling while significantly reducing cost per sample for gene and promoter capture.\n\nConclusionsWe show that a capture design employing an island strategy can enable analysis of the large gene/promoter space of wheat with only 2x160 Mb probe sets. Furthermore, these assays extend the regions of the wheat genome that are amenable to analyses beyond its exome, providing tools for detailed characterization of these regulatory regions in large populations.

genomics↗

bioRxiv2017

Epigenomic variation across a polyploid wheat diversity collection

Wheat has been domesticated into a large number of agricultural environments and has a remarkable ability to adapt to diverse environments. To understand this process, we survey genotype, repeat content and DNA methylation across a bread wheat landrace collection representing global genetic diversity. We identify independent variation in methylation, genotype and transposon copy number. We show that these, so far unexploited, sources of variation have had a massive impact on the wheat genome and that ancestral methylation states become preferentially hard coded as SNPs via 5-methylcytosine deamination. These mechanisms also drive local adaption, impacting important traits such as heading date and salt tolerance. Methylation and transposon diversity could therefore be used alongside single nucleotide polymorphism (SNP) based markers for breeding.

genomics↗

bioRxiv2017

A modified sequence capture approach allowing standard and methylation analyses of the same enriched genomic DNA sample

BackgroundBread wheat has a large complex genome that makes whole genome resequencing costly. Therefore, genome complexity reduction techniques such as sequence capture make re-sequencing cost effective. With a high-quality draft wheat genome now available it is possible to design capture probe sets and to use them to accurately genotype and anchor SNPs to the genome. Furthermore, in addition to genetic variation, epigenetic variation provides a source of natural variation contributing to changes in gene expression and phenotype that can be profiled at the base pair level using sequence capture coupled with bisulphite treatment. Here, we present a new 12 Mbp wheat capture probe set, that allows both the profiling of genotype and methylation from the same DNA sample. Furthermore, we present a method, based on Agilent SureSelect Methyl-Seq, that will use a single capture assay as a starting point to allow both DNA sequencing and methyl-seq.\n\nResultsOur method uses a single capture assay that is sequentially split and used for both DNA sequencing and methyl-seq. The resultant genotype and epi-type data is highly comparable in terms of coverage and SNP/methylation site identification to that generated from separate captures for DNA sequencing and methyl-seq. Furthermore, by defining SNP frequencies in a diverse landrace from the Watkins collection we highlight the importance of having genotype data to prevent false positive methylation calls. Finally, we present the design of a new 12 Mbp wheat capture and demonstrate its successful application to re-sequence wheat.\n\nConclusionWe present a cost-effective method for performing both DNA sequencing and methyl-seq from a single capture reaction thus reducing reagent costs, sample preparation time and DNA requirements for these complementary analyses.

genomics↗