Biology subjects

Labagnara, M.

Publications and source records attributed to Labagnara, M..

4 recordsLinked to original sources

bioRxiv2026

Love-thy-neighbor: Neural networks for tracking and lineage tracing in budding yeast

Tracking and lineage tracing are widely needed tasks in biological image analysis. For cells that grow and divide, tracking is challenging because cells change in number, shape, and size throughout a recording. As the time interval between images increases, it becomes more difficult to establish correspondences between cells across timepoints. Consequently, tracking has to be performed between consecutive or temporally close images, which leads to exponentially decreasing tracking accuracy and thus high sensitivity to error rates. For budding yeast, this challenge is further heightened by the similarity of cells in colonies, their dense packing, the asymmetric nature of cell divisions, and movement due to growth of the colony. A related task, lineage tracing, is similarly challenging without fluorescent markers due to multiple potential mother cells surrounding a new daughter cell. Here, we present neural networks for budding yeast tracking and lineage tracing, named LYN-track and LYN-trace, respectively. These methods leverage fine geometric features of cells and their neighborhoods. To train and test the algorithms, we recorded and annotated new budding and fission yeast microscopy movies (78,852 frame-to-frame tracklets, 2,512 images), which we make freely available. On these and existing datasets, our neural network-based methods demonstrate robust, above state-of-the-art performance. Both tools have been integrated into graphical user interfaces (GUIs), available on Github, and can be straightforwardly retrained with custom data if desired.

bioinformatics↗

bioRxiv2025

Optogenetic tripwires resolve models of how cells count DNA breaks

DNA end resection is a critical step in DNA damage repair and activation of the DNA damage checkpoint (DDC). To date, resection has been studied primarily in bulk cell populations. However, single-cell analyses are essential for uncovering cell-to-cell variability and can powerfully support or contradict system-level models where traditional genetic perturbations face limitations. We present a single-cell method to quantify resection by integrating an optogenetic expression system at defined distances from an inducible double-strand DNA break (DSB) site in the budding yeast genome. Using this system, we test competing, unresolved models of how the DDC counts DSBs and determines when to override the checkpoint. Current models propose that the extent of DNA damage is signaled by resection, either through liberated single-stranded DNA (ssDNA) or proteins bound along the non-resected strand. Although mechanistically plausible and widely known, these models rely on inconclusive evidence from gene knockout studies. An alternative hypothesis is that the DDC counts DSBs digitally, using factors located at 3 break ends or at ss/dsDNA junctions. Here, we leverage natural cell-to-cell variability in resection rates as an intrinsic perturbation, avoiding the limitations of prior genetic approaches. Our single-cell data challenge models in which DNA damage is inferred from the extent or rate of resection or from proteins bound along resected DNA. To explore alternative mechanisms, we investigated DDC proteins localized at 3 DSB ends or ss/dsDNA boundaries. By dynamically depleting candidate proteins after checkpoint arrest, we identified ss/dsDNA boundary proteins Ddc1, Dpb11, and Rad9 as essential for DDC maintenance and promising candidates for a cascade that acts as a digital DSB counter. Our findings demonstrate that quantitative, system-level single-cell approaches, coupled with dynamic perturbations, can resolve fundamental questions in DNA repair and checkpoint signaling.

molecular biology↗

bioRxiv2024

Automated plasmid design for marker-free genome editing in budding yeast

The ease of genome editing has contributed to the popularity of budding yeast as a model organism. However, the palette of selectable markers is in principle limited as most can only be used once. Some markers such as URA3 and TRP1 can be recycled through counterselection. This permits seamless genome modification with pop-in/pop-out (PIPO), in which a DNA construct first integrates in the genome and, subsequently, homologous regions recombine and excise undesired sequences. Popular approaches for creating such constructs use oligonucleotides and polymerase chain reaction (PCR). The drawbacks are that long oligonucleotides are unstable, can form secondary structures that interfere with PCR, cannot be regenerated in a typical biological laboratory, and are only widely available for lengths less than about 120, which limits the homology and efficiency that can be attained. With the rapid reduction in price, synthesizing custom DNA sequences in specific plasmid backbones has become an appealing alternative. For designing plasmids for seamless PIPO gene tagging or deletion, there are a number of factors to consider. To create only the shortest DNA sequences necessary, avoid errors in manual design, specify the amount of homology desired, and customize restriction sites, we created the computational tool PIPOline. Using it, we tested the ratios of homology that improve pop-out efficiency when targeting the genes HTB2 or WHI5. We supply optimal PIPO plasmid sequences for tagging or deleting almost all S288C budding yeast open reading frames (ORFs). Finally, we demonstrate how the histone variant Htb2 marked with a red fluorescent protein can be used as a cell-cycle stage marker, alternative to superfolder GFP (sfGPF), reducing light toxicity. We expect PIPOline to streamline genome editing in budding yeast.

bioinformatics↗

bioRxiv2024

Light-directed evolution of dynamic, multi-state, and computational protein functionalities

Directed evolution is a powerful method in biological engineering. Current approaches draw on time-invariant selection mechanisms, ideal for evolving steady-state properties such as enzymatic activity or fluorescence intensity. A fundamental problem remains how to continuously evolve dynamic, multi-state, or computational functionalities, e.g., on-off kinetics, state-specific activity, stimulus-responsiveness, or switching and logic capabilities. These require selection pressure on all of the states of a protein of interest (POI) and the transitions between them. We realized that optogenetics and cell cycle oscillations could be leveraged for a novel directed evolution paradigm ( optovolution) that is germane for this need: We designed a signaling cascade in budding yeast where optogenetic input switches the POI between off (0) and on (1) states. In turn, the POI controls a Cdk1 cyclin, which in the re-engineered cell cycle system is essential for one cell cycle stage but poisonous for another. Thus, the cyclin must oscillate (1-0-1-0...) for cell proliferation. In this system, evolution can act efficiently on the POIs different states, input-output relations, and dynamics on the timescale of minutes in every cell cycle. Further, controlling the pacemaker, light, directs and tunes selection pressures. Optovolution is in vivo, continuous, self-selecting, and efficient. We first evolved two optogenetic systems, which relay 0/1 input to 0/1 output: We obtained 19 new variants of the LOV transcription factor El222 that were stronger, less leaky, or green light responsive in vivo. We demonstrate the utility of the latter mutations for orthogonal color-multiplexing with only LOV domains for the first time. Evolving the PhyB-Pif3 optogenetic system, we discovered that loss of YOR1 makes supplementing the chromophore phycocyanobilin (PCB) unnecessary. Finally, we demonstrate the generality of the method by evolving a destabilized rtTA transcription factor, which performs an AND operation between transcriptional and doxycycline input. Optovolution makes coveted, difficult-to-change protein functionalities continuously evolvable.

synthetic biology↗