Johan Elf

This paper develops a vision transformer model for rapid bacterial species identification from time-lapse images without fluorescent labels, achieving high accuracy on clinical isolates.

The paper uses optical pooled screening to analyze a bacterial transposon mutagenesis library via live cell imaging, identifying mutants with replication initiation control perturbations in E. coli.

This paper develops a method to rapidly detect bacteria in blood using smart centrifugation, microfluidic trapping, and deep learning for sepsis diagnostics.

The paper develops a method for in situ genotyping of bacterial libraries using chromosomally expressed barcodes to measure intracellular maturation times of red fluorescent proteins.

This study uses single-cell imaging and a deep neural network to rapidly assess drug susceptibility in Mycobacterium tuberculosis variant bovis BCG, achieving results in under 12 hours.

This study investigates how mutations in the hinge region of LacI affect binding stability and inducibility, showing a trade-off between binding strength and inducibility rather than speed and stability.

This study evaluates if phenotypic antibiotic susceptibility testing can be performed on single bacterial cells by monitoring phenotypic changes over time, finding that single-cell responses are heterogeneous and require sequential antibiotic exposure for accurate resistance classification.

The paper measures association and dissociation rates of LacI binding sites in living cells and finds an anti-correlation between these rates, challenging the assumption that binding site strength is solely determined by dissociation rate.

This study uses live-cell imaging of fluorescently labeled loci in E. coli to inform a dynamic polymer model of chromosome organization across the cell cycle, revealing global reorganization of the chromosome structure.

This paper shows that the 3D point spread function of a fluorescent emitter can be directly accessed in the latent space of an autoencoder trained on sample images without z-reference data.

The paper uses deep learning to classify bacteria species from phase-contrast microscopy images without staining or fixation, achieving over 98% accuracy.

This paper develops a method for accurate 3D localization of fluorescent proteins in living E. coli using mother-machine devices and deep learning, revealing distinct nucleoid locus distributions.

This paper introduces a method for real-time pooled optical screening with single-cell isolation using optical tweezers and 3D microfluidics.

This study directly measures mRNA translation kinetics in living cells using a live-cell ribosome-labeling method and single-molecule tracking.

This study measures the rotation-coupled diffusion of LacI along DNA, finding it moves ~50 base pairs per revolution, supporting a facilitated diffusion model.

The paper introduces variational algorithms for analyzing noisy multi-state diffusion trajectories in single particle tracking data, improving speed and supporting broader inference methods.

The paper demonstrates in situ genotyping of pooled strain libraries by combining single molecule fluorescence imaging with barcoded plasmids to identify individual cell genotypes after phenotype characterization.

The paper presents a fast antibiotic susceptibility testing method that determines bacterial growth rate responses to antibiotics in under 25 minutes using microfluidic imaging.