Luke Evans · University of Colorado Boulder · PhdFit
Resume-aware faculty matching
Find professors who actually fit you
Upload your resume. PhdFit's six research agents compare your background with faculty profiles, recent publications, lab focus, and outreach opportunities, then rank professors with evidence you can review.
Resume-aware matchingEvidence from recent researchOutreach drafts included
Advisor Match Brief Resume uploaded
Resume signals5 found
NLPHCIFairnessGraph MLPyTorch
Six-agent analysisRunning
Axel · Candidate AnalystResume parsed
Nova · Professor ResearcherRecent papers
Echo · Match ExplainerEvidence linked
Atlas · Strategy AdvisorFit ranked
Lyra · Outreach WriterDraft queued
Juno · Meeting Prep CompilerNext step ready
SC
Dr. Sarah Chen
Stanford · NLP & interpretability
91
fit
Fairness + NLP projects overlap with her recent work.
Strong methods fit: model evaluation and interpretability.
Clear outreach angle for current trustworthy AI research.
Outreach angle
Ask how her lab is extending interpretability methods into fairness audits for real-world AI systems.
Luke Evans
· Assistant ProfessorVerified
University of Colorado Boulder · Ecology & Evolutionary Biology
Luke Evans is an Assistant Professor in the Department of Ecology and Evolutionary Biology at the University of Colorado Boulder. He earned his Ph.D. from Northern Arizona University in 2012. His research focuses on methods for estimating heritability and investigating the genetic architecture of complex traits, with recent work emphasizing human psychiatric disorders. This includes identifying genetic variants that contribute to disorder liability, understanding gene-by-environment interactions, and testing how these factors influence risk prediction. Additionally, he studies selection and adaptation by identifying quantitative and molecular signatures of selection, both recent positive and purifying selection. His work involves large-scale reciprocal transplant experiments to test gene-by-environment interaction hypotheses and to understand how environments shape complex traits, particularly in ecologically and economically important tree genera such as Populus and Salix.
Many datasets, including widely used biobanks, have more than one observation of numerous phenotypes for at least a portion of their sample. The majority of genome-wide association studies (GWASs) utilize only a single observation per individual, even when more than one observation may be available, and apply a standard model in which the additive allelic effect being estimated is assumed to be constant across the age or time range in the sample. Here, we test a set of simple approaches to utilize multiple observations per individual, under this same assumption, to characterize effects on GWAS power, SNP heritability, gene set enrichment, and polygenic prediction. We find that utilizing the mean or median of the available observations rather than a single observation improves the power to detect associated loci and enriched gene sets and yields higher out-of-sample polygenic score prediction accuracy. Despite growing biobanks, many deeply phenotyped samples are relatively small but have multiple observations. While explicitly modeling age- or time-dependent genetic effects can add nuance to genetic studies and estimates, most GWASs apply a standard, additive-only model; a simple approach of using the mean or median can improve power by reducing "noise" in the phenotype, utilize standard, optimized software, and be particularly impactful for smaller samples, including samples of diverse genetic ancestry existing in widely used biobanks such as the UK Biobank and the Health and Retirement Study.
Thyroid diseases are common and highly heritable. We performed a meta-analysis of genome-wide association studies from 19 biobanks for five thyroid diseases: thyroid cancer (ThC), benign nodular goiter, Graves' disease, lymphocytic thyroiditis and primary hypothyroidism. We analyzed genetic association data from ~2.9 million genomes and identified 313 known and 570 new independent loci linked to thyroid diseases. We discovered genetic correlations between ThC, benign nodular goiter and autoimmune thyroid diseases (rg = 0.16-0.97). Telomere maintenance genes contributed to benign and malignant thyroid nodular disease risk, whereas cell cycle, DNA repair and damage response genes were associated with ThC. We propose a paradigm that explains genetic predisposition to benign and malignant thyroid nodules. We found polygenic risk score associations with ThC risk of structural disease recurrence, tumor size, multifocality, lymph node metastases and extranodal extension. Polygenic risk scores identified individuals with aggressive ThC in a biobank, creating an opportunity for genetically informed population screening.
= 0.18). Next-generation PCGrimAge and DunedinPace clocks showed consistent zygosity correlations across tissues, while multi-tissue clocks (e.g. ZhangQ) showed comparable MZ-DZ correlations. While saliva-based DNA methylation is not a direct substitute for blood-based DNA methylation, BC and PBMC show comparability; nevertheless, all tissue types may be appropriate for DNA methylation aging studies when compared within tissues.
Epigenetic aging prior to midlife is gaining interest as an intervenable period to address health and cognitive aging. Epigenetic changes may index DNA methylation aging rates, but methylation profiles may not be substitutable across tissues. We compared DNA methylation clocks and age acceleration in saliva, buffy coat (BC), and peripheral blood mononuclear cells (PBMC) collected in 91 individuals (7 unpaired, 20 siblings, 64 twins; 18 monozygotic (MZ), 14 dizygotic (DZ) pairs) from the Colorado Adoption/Twin Study of Lifespan behavioral development and cognitive aging (CATSLife1; Mean age = 30.90 years [range = 28.07–41.13]; 50.5% female). Across 15 DNA methylation clocks, chronological age and DNA methylation ages were moderately associated (mean Spearman correlations: <i>r</i> = 0.37, Saliva; <i>r</i> = 0.40, BC; <i>r</i> = 0.34, PBMC). In mixed-effects models, saliva showed higher DNA methylation ages (<i>B</i> = 3.77–19.72 years vs BC, <i>p</i> < 0.001), whereas PBMC and BC were comparable (<i>B</i> = −0.06–0.39 years vs BC, <i>p</i> ≥ 0.436). The exception was the next-generation clock DunedinPace showing comparability (<i>p</i> = 0.486). Similar patterns were observed for age acceleration estimates. Altogether, MZ pairs (meta-analytic <i>r =</i> 0.49, 95%CI = 0.32,0.66) and DZ pairs (meta-analytic <i>r =</i> 0.38, 95%CI = 0.25,0.52) were moderately correlated (Spearman), but MZ pairs showed heterogeneity across tissues (<i>p</i> < 0.034): saliva was lower (mean <i>r =</i> 0.33, <i>SD</i> = 0.25) than BC (mean <i>r =</i> 0.64, <i>SD</i> = 0.10) and PBMC (mean <i>r =</i> 0.49, <i>SD</i> = 0.18). Next-generation PCGrimAge and DunedinPace clocks showed consistent zygosity correlations across tissues, while multi-tissue clocks (e.g. ZhangQ) showed comparable MZ-DZ correlations. While saliva-based DNA methylation is not a direct substitute for blood-based DNA methylation, BC and PBMC show comparability; nevertheless, all tissue types may be appropriate for DNA methylation aging studies when compared within tissues. Saliva-based methylation is not a direct substitute for blood-based methylation.Saliva-based DNA methylation clocks demonstrated relatively older methylation ages and faster age accelerations relative to blood-based DNA methylation clocks.Blood-based DNA methylation clocks were comparable across buffy coat and peripheral blood mononuclear cell tissues.Twin similarity, particularly among identical twins, was higher for blood-based DNA methylation clocks and newer generation DNA methylation clocks.Three clocks (one first-generation multi-tissue and two next-generation) showed consistency and/or comparability across tissues and showed moderate to strong correlations among identical and fraternal twins, supporting the importance of tool development. Saliva-based methylation is not a direct substitute for blood-based methylation. Saliva-based DNA methylation clocks demonstrated relatively older methylation ages and faster age accelerations relative to blood-based DNA methylation clocks. Blood-based DNA methylation clocks were comparable across buffy coat and peripheral blood mononuclear cell tissues. Twin similarity, particularly among identical twins, was higher for blood-based DNA methylation clocks and newer generation DNA methylation clocks. Three clocks (one first-generation multi-tissue and two next-generation) showed consistency and/or comparability across tissues and showed moderate to strong correlations among identical and fraternal twins, supporting the importance of tool development.
Epigenetic aging prior to midlife is gaining interest as an intervenable period to address health and cognitive aging. Epigenetic changes may index DNA methylation aging rates, but methylation profiles may not be substitutable across tissues. We compared DNA methylation clocks and age acceleration in saliva, buffy coat (BC), and peripheral blood mononuclear cells (PBMC) collected in 91 individuals (7 unpaired, 20 siblings, 64 twins; 18 monozygotic (MZ), 14 dizygotic (DZ) pairs) from the Colorado Adoption/Twin Study of Lifespan behavioral development and cognitive aging (CATSLife1; Mean age = 30.90 years [range = 28.07–41.13]; 50.5% female). Across 15 DNA methylation clocks, chronological age and DNA methylation ages were moderately associated (mean Spearman correlations: <i>r</i> = 0.37, Saliva; <i>r</i> = 0.40, BC; <i>r</i> = 0.34, PBMC). In mixed-effects models, saliva showed higher DNA methylation ages (<i>B</i> = 3.77–19.72 years vs BC, <i>p</i> < 0.001), whereas PBMC and BC were comparable (<i>B</i> = −0.06–0.39 years vs BC, <i>p</i> ≥ 0.436). The exception was the next-generation clock DunedinPace showing comparability (<i>p</i> = 0.486). Similar patterns were observed for age acceleration estimates. Altogether, MZ pairs (meta-analytic <i>r =</i> 0.49, 95%CI = 0.32,0.66) and DZ pairs (meta-analytic <i>r =</i> 0.38, 95%CI = 0.25,0.52) were moderately correlated (Spearman), but MZ pairs showed heterogeneity across tissues (<i>p</i> < 0.034): saliva was lower (mean <i>r =</i> 0.33, <i>SD</i> = 0.25) than BC (mean <i>r =</i> 0.64, <i>SD</i> = 0.10) and PBMC (mean <i>r =</i> 0.49, <i>SD</i> = 0.18). Next-generation PCGrimAge and DunedinPace clocks showed consistent zygosity correlations across tissues, while multi-tissue clocks (e.g. ZhangQ) showed comparable MZ-DZ correlations. While saliva-based DNA methylation is not a direct substitute for blood-based DNA methylation, BC and PBMC show comparability; nevertheless, all tissue types may be appropriate for DNA methylation aging studies when compared within tissues. Saliva-based methylation is not a direct substitute for blood-based methylation.Saliva-based DNA methylation clocks demonstrated relatively older methylation ages and faster age accelerations relative to blood-based DNA methylation clocks.Blood-based DNA methylation clocks were comparable across buffy coat and peripheral blood mononuclear cell tissues.Twin similarity, particularly among identical twins, was higher for blood-based DNA methylation clocks and newer generation DNA methylation clocks.Three clocks (one first-generation multi-tissue and two next-generation) showed consistency and/or comparability across tissues and showed moderate to strong correlations among identical and fraternal twins, supporting the importance of tool development. Saliva-based methylation is not a direct substitute for blood-based methylation. Saliva-based DNA methylation clocks demonstrated relatively older methylation ages and faster age accelerations relative to blood-based DNA methylation clocks. Blood-based DNA methylation clocks were comparable across buffy coat and peripheral blood mononuclear cell tissues. Twin similarity, particularly among identical twins, was higher for blood-based DNA methylation clocks and newer generation DNA methylation clocks. Three clocks (one first-generation multi-tissue and two next-generation) showed consistency and/or comparability across tissues and showed moderate to strong correlations among identical and fraternal twins, supporting the importance of tool development.
BACKGROUND: Frailty, an age-related loss of the ability to withstand stressors, is commonly measured using health deficit indices, often using survey or questionnaire data. We aimed to develop an electronic frailty index (eFI) using electronic health record (EHR) data linkages in the UK Biobank and assess its association with mortality. METHODS: We calculated an eFI using 43 deficits, each corresponding to phecodes mapped to the United Kingdom (UK) and international classification coding systems. We compared this eFI to a validated 49-item survey-based FI for the UK Biobank and assessed associations of the eFI with risk of all-cause mortality (follow-up ≤ 10.2 years) and mortality following a stressor (heart attack or stroke) using Cox proportional hazard models. RESULTS: Mean eFI in this cohort (N = 208,982) was 0.058 (SD = 0.06) and was higher in females than males. A 10% higher baseline frailty was associated with higher risk of all-cause mortality (HR(95%CI)=2.00(1.93-2.07)), although the magnitude of this association decreased when adjusting for socioeconomic-related covariates (HR(95%CI)=1.44(1.38-1.51)). Associations were stronger in men than women. eFI predicted mortality following both heart attack and stroke (HR(95%CI)=1.59(1.25-2.04) and HR = 1.33(1.13-1.57), respectively). CONCLUSIONS: This EHR-based eFI has robust associations with mortality, suggesting that it can be used as a valid measure of frailty in the UK Biobank and can potentially be applied to other datasets with EHR data.
bioRxiv (Cold Spring Harbor Laboratory) · 2025-09-18
preprintOpen access
Abstract Epigenetic aging prior to midlife is gaining interest as an intervenable period to address health and cognitive aging. Epigenetic changes may index DNA methylation aging rates, but methylation profiles may not be substitutable across tissues. We compared DNA methylation clocks and age acceleration in saliva, buffy coat (BC), and peripheral blood mononuclear cells (PBMC) collected in 91 individuals (7 unpaired, 20 siblings, 64 twins; 18 monozygotic (MZ), 14 dizygotic (DZ) pairs) from the Colorado Adoption/Twin Study of Lifespan behavioral development and cognitive aging (CATSLife1; Mean age=30.90 years [range=28.07–41.13]; 50.5% female). Across 15 DNA methylation clocks, chronological age and DNA methylation ages were moderately associated (mean Spearman correlations: r =0.37, Saliva; r =0.41, BC; r =0.38, PBMC). In mixed-effects models, saliva showed higher DNA methylation ages ( B =3.83–16.46 years vs BC, p<0.001), whereas PBMC and BC were comparable ( B =-0.06–0.39 years vs BC, p > 0.447). The exception was the next-generation clock DunedinPace showing comparability ( p =0.486). Similar patterns were observed for age acceleration estimates. Altogether, MZ pairs (meta-analytic r= 0.49, 95%CI=0.32,0.66) and DZ pairs (meta-analytic r= 0.38, 95%CI=0.25,0.52) were moderately correlated (Spearman), but MZ pairs showed heterogeneity across tissues ( p <0.020): saliva was lower (mean r= 0.33, SD =0.25) than BC (mean r= 0.64, SD =0.10) and PBMC (mean r= 0.49, SD =0.18). Next-generation PCGrimAge and DunedinPace clocks showed consistent zygosity correlations across tissues, while multi-tissue clocks (e.g., ZhangQ) showed comparable MZ-DZ correlations. While saliva-based DNA methylation is not a direct substitute for blood-based DNA methylation, BC and PBMC show comparability; nevertheless, all tissue types may be appropriate for DNA methylation aging studies when compared within tissues. Key Policy Highlights Saliva-based methylation is not a direct substitute for blood-based methylation. Saliva-based DNA methylation clocks demonstrated relatively older methylation ages and faster age accelerations relative to blood-based DNA methylation clocks. Blood-based DNA methylation clocks were comparable across buffy coat and peripheral blood mononuclear cell tissues. Twin similarity, particularly among identical twins, was higher for blood-based DNA methylation clocks and newer generation DNA methylation clocks. Three clocks (one first-generation multi-tissue and two next-generation) showed consistency and/or comparability across tissues and showed moderate to strong correlations among identical and fraternal twins, supporting the importance of tool development.
Journal of Chemical Theory and Computation · 2025-12-01 · 1 citations
article
Markov state models (MSMs) are widely employed to analyze the kinetics of complex systems. But despite their effectiveness in many applications, MSMs are prone to systematic or statistical errors, often exacerbated by suboptimal hyperparameter choice. In this article, we attempt to understand how these choices affect the error of estimates of mean first-passage times and committors, key quantities in chemical rate theory. We first evaluate the performance of the recently introduced “stopped-process estimator” [Strahan, J. Long-time-scale predictions from short-trajectory data: A benchmark analysis of the trp-cage miniprotein. J. Chem. Theory Comput. 2021, 17, 2948–2963. 10.1021/acs.jctc.0c00933.] that attempts to reduce error caused by choosing a too-large lag time. We then study the effect of statistical errors on Markov state model construction using the condition number, which measures an MSM’s sensitivity to perturbation. This analysis helps give an insight into which factors cause an MSM to be more or less sensitive to statistical error. Our work highlights the importance of choosing a good sampling measure, the measure from which the initial points are drawn, and has implications for recent work applying a variational principle for evaluating the committor.
