Biology subjects

Montgomery, G.

Publications and source records attributed to Montgomery, G..

7 recordsLinked to original sources

bioRxiv2026

Predicting Endometriosis Status and Menstrual Cycle Phase Using DNA Methylation

Endometriosis is a chronic inflammatory disease associated with pelvic pain, infertility, and delayed diagnosis. Growing evidence suggests that altered DNA methylation contributes to disease development and could serve as a biomarker for disease. We developed a leakage-safe machine learning pipeline to classify endometriosis case-control status and menstrual cycle phase using genome-wide DNA methylation data from eutopic endometrial tissue. The dataset consisted of 984 samples profiled using the Illumina Infinium MethylationEPIC array, with measurements across approximately 759,000 CpG sites. Technical variation was corrected using SmartSVA batch correction. Ridge logistic regression models were trained using stratified 80/20 train-test splits, with regularization strength selected via stratified cross-validation. Feature selection approaches included ridge coefficient ranking, per-CpG t-tests, and univariate logistic regression with FDR correction. Model validity was evaluated using label-shuffling analyses. Menstrual cycle phase classification showed strong performance (mean cross-validation AUROC: 0.971, held-out test AUROC: 0.989), reflecting genome-wide hormonally driven methylation. Ridge regression produced lower but meaningful performance for endometriosis classification (mean cross-validation AUROC: 0.854, held-out test AUROC: 0.875). Ridge coefficient-based feature selection identified compact predictive CpG sets, supporting the hypothesis that endometriosis-associated methylation signal is distributed across many loci rather than a few highly predictive CpGs. Pathway enrichment analyses identified substantial enrichment for menstrual cycle phase but limited enrichment for disease status following FDR correction, consistent with a diffuse endometriosis-associated signal. These findings demonstrate that ridge regression can detect methylation patterns associated with both endometriosis and menstrual cycle phase, highlighting the importance of accounting for cycle-related epigenetic variation in endometrial DNA methylation studies.

bioinformatics↗

bioRxiv2021

An extremes of phenotype approach confirms significant genetic heterogeneity in patients with ulcerative colitis.

Ulcerative colitis (UC) is a major form of inflammatory bowel disease with increasing global incidence. There is significant phenotypic heterogeneity defined by a range of clinical variables including age of onset and disease extent. Clinical outcomes range from long-term remission on minimal therapy to surgical resection. Close to 70% of UC risk can be attributed to genetics and understanding the genetic mechanisms contributing to this risk and disease heterogeneity is vital for understanding disease pathogenesis and improving patient outcomes through targeted screening and therapies. This study aims to characterise the genetic heterogeneity of UC by identifying genomic risk variants specific to mild and/or severe forms of UC, exploring variations in the effect size of known risk variants and assessing the clinical value of a genetic risk score (GRS). We conducted genome-wide association (GWA) analyses in 287 patients with mild UC, 311 patients with severe UC and 583 age- and gender-matched controls. Odds ratios (OR) for mild vs control, severe vs control and combined mild and severe UC vs control were calculated. Using the combined UC data, two independent loci in the HLA region reached genome-wide significance. An additional genome-wide significant signal on chromosome 1 was identified in severe cases only. OR for known risk loci varied between mild and severe patients and were similar to previously published results. Effect estimates from the most recent UC GWA meta-analysis were used to calculate a GRS for each individual. A higher mean GRS was observed in both mild and severe UC cases compared to controls however, there was no difference between the mean GRS for mild and severe UC. Heterogeneity in effect sizes of UC associated variants between mild and severe disease burden suggests the presence of genetically distinct signatures. While large consortium data are needed to identify genome-wide significant variants, additional risk loci may be identified by targeted recruitment of individuals with a history of severe disease. Author SummaryUlcerative colitis (UC) is a chronic and often debilitating form of inflammatory bowel disease affecting approximately 0.3% of the population in industrialized economies. The disease displays significant clinical heterogeneity including age at presentation, disease severity, and the propensity to develop disease-related complications. Several previous studies have demonstrated the heritability of UC, identifying over 30 loci specific to the disease. The majority of these loci have small to modest effect sizes other than those within the Human Leucocyte Antigen (HLA) region on chromosome 6. Using stringent clinical criteria for defining mild and severe forms of UC in an extremes of phenotype approach, we undertook a genome wide association study in a dataset of 1222 participants to investigate genetic heterogeneity in this disease. We demonstrated substantial differences in genetic associations in severe UC as compared to mild UC. While over 2,000 SNPs achieved genome-wide significance in the severe UC analysis, none reached significance for mild UC. These results were reflected in significant differences in odds ratios. We identified Complement Factor B (CFB) as a potential susceptibility gene for severe UC in the Caucasian population with additional tissue gene expression demonstrating a positive correlation with disease severity.

genetics↗

bioRxiv2021

Enhanced Solid Tumor Recognition and T cell Stemness with SynNotch CAR Circuits

The lack of highly tumor-specific antigens limits the development of engineered T cell therapeutics because of life-threatening "on-target/off-tumor" toxicities. Here we identify ALPPL2 as a tumor-specific antigen expressed in a spectrum of solid tumors, including mesothelioma. ALPPL2 can act as a sole target for chimeric antigen receptor (CAR) therapy or be combined with tumor-associated antigens such as MCAM or mesothelin in synthetic Notch (synNotch) CAR combinatorial antigen circuits. SynNotch CAR T cells display superior tumor control when compared to CAR T cells to the same antigens by prevention of CAR-mediated tonic signaling allowing T cells to maintain a long-lived memory and non-exhausted phenotype. Collectively, we establish ALPPL2 as a clinically viable target for multiple solid tumors and demonstrate the multi-faceted therapeutic benefits of synNotch CAR T cells. ONE SENTENCE SUMMARYSynNotch CAR circuits targeting novel solid tumor antigens enhance specificity and improve therapeutic efficacy by regulating T cell exhaustion.

immunology↗

bioRxiv2018

Genome Scale Epigenetic Profiling Reveals Five Distinct Subtypes of Colorectal Cancer

BACKGROUNDColorectal cancer is an epigenetically heterogeneous disease, however the extent and spectrum of the CpG Island Methylator Phenotype (CIMP) is not clear.\n\nRESULTSAn unselected cohort of 216 colorectal cancers clustered into five clinically and molecularly distinct subgroups using Illumina 450K DNA methylation arrays. CIMP-High cancers were most frequent in the proximal colons of female patients. These dichotomised into CIMP-Hl and CIMP-H2 based on methylation profile which was supported by over representation of BRAF (74%, P<0.0001) or KRAS (55%, P<0.0001) mutation, respectively. Congruent with increasing methylation, there was a stepwise increase in patient age from 62 years in the CI MP-Negative subgroup to 75 years in the CIMP-Hl subgroup (P<0.0001). There was a striking association between PRC2-marked loci and those subjected to significant gene body methylation in CIMP-type cancers (P<1.6xl078). We identified oncogenes susceptible to gene body methylation and Wnt pathway antagonists resistant to gene body methylation. CIMP cluster specific mutations were observed for genes involved in chromatin remodelling, such as in the SWI/SNF and NuRD complexes, suggesting synthetic lethality.\n\nCONCLUSIONThere are five clinically and molecularly distinct subgroups of colorectal cancer based on genome wide epigenetic profiling. These analyses highlighted an unidentified role for gene body methylation in progression of serrated neoplasia. Subgroup-specific mutation of distinct epigenetic regulator genes revealed potentially druggable vulnerabilities for these cancers, which may provide novel precision medicine approaches.

genomics↗

bioRxiv2018

Improved prediction of chronological age from DNA methylation limits it as a biomarker of ageing

DNA methylation is associated with age. The deviation of age predicted from DNA methylation from actual age has been proposed as a biomarker for ageing. However, a better prediction of chronological age implies less opportunity for biological age. Here we used 13,661 samples (from blood and saliva) in the age range of 2 to 104 years from 14 cohorts measured on Illumina HumanMethylation450/EPIC arrays to perform prediction analyses. We show that increasing the sample size achieves a smaller prediction error and higher correlations in test datasets. We demonstrate that smaller prediction errors provide a limit to how much variation in biological ageing can be captured by methylation and provide evidence that age predictors from small samples are prone to confounding by cell composition. Our predictor shows a similar or better performance in non-blood tissues including saliva, endometrium, breast, liver, adipose and muscle, compared with Horvaths across-tissue age predictor.

bioinformatics↗

bioRxiv2017

The Anorexia Nervosa Genetics Initiative: Overview and Methods

BackgroundGenetic factors contribute to anorexia nervosa (AN); and the first genome-wide significant locus has been identified. We describe methods and procedures for the Anorexia Nervosa Genetics Initiative (ANGI), an international collaboration designed to rapidly recruit 13000 individuals with AN as well as ancestrally matched controls. We present sample characteristics and the utility of an online eating disorder diagnostic questionnaire suitable for large-scale genetic and population research.\n\nMethodsANGI recruited from the United States (US), Australia/New Zealand (ANZ), Sweden (SE), and Denmark (DK). Recruitment was via national registers (SE, DK); treatment centers (US, ANZ, SE, DK); and social and traditional media (US, ANZ, SE). All cases had a lifetime AN diagnosis based on DSM-IV or ICD-10 criteria (excluding amenorrhea). Recruited controls had no lifetime history of disordered eating behaviors. To assess the positive and negative predictive validity of the online eating disorder questionnaire (ED100K-v1), 109 women also completed the Structured Clinical Interview for DSM-IV (SCID), Module H.\n\nResultsBlood samples and clinical information were collected from 13,364 individuals with lifetime AN and from controls. Online diagnostic phenotyping was effective and efficient; the validity of the questionnaire was acceptable.\n\nConclusionsOur multipronged recruitment approach was highly effective for rapid recruitment and can be used as a model for efforts by other groups. High online presence of individuals with AN rendered the Internet/social media a remarkably effective recruitment tool in some countries. ANGI has substantially augmented Psychiatric Genomics Consortium AN sample collection. ANGI is a registered clinical trial: clinicaltrials.gov NCT01916538; https://clinicaltrials.gov/ct2/show/NCT01916538?cond=Anorexia+Nervosa&draw=1&rank=3.

genetics↗

bioRxiv2017

Widespread signatures of negative selection in the genetic architecture of human complex traits

Estimation of the joint distribution of effect size and minor allele frequency (MAF) for genetic variants is important for understanding the genetic basis of complex trait variation and can be used to detect signature of natural selection. We develop a Bayesian mixed linear model that simultaneously estimates SNP-based heritability, polygenicity (i.e. the proportion of SNPs with nonzero effects) and the relationship between effect size and MAF for complex traits in conventionally unrelated individuals using genome-wide SNP data. We apply the method to 28 complex traits in the UK Biobank data (N = 126,752), and show that on average across 28 traits, 6% of SNPs have nonzero effects, which in total explain 22% of phenotypic variance. We detect significant (p < 0.05/28 =1.8x10-3) signatures of natural selection for 23 out of 28 traits including reproductive, cardiovascular, and anthropometric traits, as well as educational attainment. We further apply the method to 27,869 gene expression traits (N = 1,748), and identify 30 genes that show significant (p < 2.3x10-6) evidence of natural selection. All the significant estimates of the relationship between effect size and MAF in either complex traits or gene expression traits are consistent with a model of negative selection, as confirmed by forward simulation. We conclude that natural selection acts pervasively on human complex traits shaping genetic variation in the form of negative selection.

genetics↗