Carl Stone

Carl Stone

Microbial evolution researcher finishing a PhD at Vanderbilt University. Expert in functional genomics and bacterial epigenetics.

Research

My work asks how bacteria integrate molecular information across scales — from individual regulatory sites to genome-wide epigenetic patterns, from single growth curves to evolutionary trajectories spanning thousands of generations. I study this using high-throughput fitness assays, genome-scale methylation analysis, and Bayesian statistical modeling.

The E. coli Dam Methylome

Dam methylase deposits N6-methyladenine (6mA) at GATC motifs throughout the E. coli genome. Bulk methylation approaches suggest near-complete modification, but I produced the most detailed per-site characterization of this methylome to date. Using long-read sequencing on experimental evolution lines, I identified 177 double-stranded GATC sites that are reproducibly hypomethylated — and found that these cluster specifically at regulatory DNA-protein interaction sites: −35 promoter elements, CRP/Fis/Fnr binding motifs, cryptic prophages, and IS elements. The absolute differences are small (~92% vs. 97%), but methylation state at a single promoter is binary, with well-documented phenotypic consequences in phase-variable regulatory systems.

I also characterized how methylation diverges in MMR-deficient evolved clones, finding hypomethylation at more sites, with greater variability and non-deterministic divergence, concentrated at sites already predisposed to low methylation. Together, this work reveals the E. coli methylome as spatially organized and functionally constrained, not merely a saturation artifact.

Publication: Stone CJ, Boyer GF, Behringer MG. 2023. Differential adenine methylation analysis reveals increased variability in 6mA in the absence of methyl-directed mismatch repair. mBio 14, e01289-23. Link

Software: commaKit — R package for differential methylation analysis, developed alongside this work

Bayesian Longitudinal Fitness Analysis and the Fitness Seascape

Standard RB-TnSeq fitness assays compare two timepoints, collapsing temporal dynamics into a single endpoint estimate. I built a Bayesian hierarchical model (Stan/CmdStanR) for longitudinal Tn-Seq that estimates time-resolved selection rates across the full E. coli growth curve — from exponential growth through long-term stationary phase (0 → 1 → 4 → 10 days). The model uses piecewise linear regression on selection rates, partially pooled across ~15 insertion mutants per gene, with correlated interval coefficients. This makes uncertain estimates conservative while allowing confidently non-zero effects to pull away from zero.

Using this framework, I constructed an empirical one-dimensional fitness seascape: a latent phenotypic axis organizing ~3,000 transposon mutant genotypes, with an optimum that moves as an OU process over time. Key findings: (1) the long-term stationary phase optimum is closer to exponential growth than to death-phase tolerance, suggesting growth capacity is the predominant selection target in LTSP; (2) axis position maps onto a ~130-condition RB-TnSeq compendium, with one end representing near-neutral generalists and the other stress-tolerant but slow-growing alleles; (3) axis position predicts evolutionary arrival time in an independent LTEP — “growth” alleles fix ~50–100 days earlier.

Preprint: Stone CJ, Behringer MG. 2026. Bayesian analysis of longitudinal RB-TnSeq resolves the fitness seascape in fluctuating environments. bioRxiv. Link (in review at Molecular Biology and Evolution)

commaKit — R Package for Bacterial Methylation Analysis

commaKit is an R package for differential bacterial DNA methylation analysis, developed because no adequate tool existed for per-site 6mA analysis in bacteria. Built alongside the mBio 2023 paper, it has since been substantially rewritten with full test coverage, vignettes, and documentation. Bioconductor submission is pending; a companion methods paper is in preparation for Microbial Resource Announcements.

Earlier Work

Virulence regulation in Staphylococcus aureus (Georgetown, 2018–2020). As a research assistant in the Brinsmade Lab, I studied how branched-chain fatty acids modulate the Sae two-component system. Publications: Pendleton et al., mBio 2022; Mlynek et al., J. Bacteriology 2020.

Microbial ecology tools benchmarking (U. Minnesota, 2017–2018). Undergraduate honors thesis comparing mothur, QIIME, and DADA2 on soil microbiome datasets. Also contributed to: Behringer et al., mBio 2022.