Polyploid cells, which contain more than two pairs of homologous chromosomes, are widely present across protozoa, plants and animals. Polyploidization occurs in many cell and tissue types as part of a developmental program, and is crucial for increases in metabolic output, cell mass and cell size. Two cell cycle variations, endoreplication and endomitosis, are known to give rise to polyploid cells. Although the function and molecular regulation of endoreplication has been extensively studied, the molecular factors that control endomitosis are largely unknown. Specifically, it is unclear how endomitotic cycles are initiated and executed, and what the advantages are of endomitosis over endoreplication. The C. elegans intestine provides an ideal system to study these questions, as transitions between canonical, endomitotic and endoreplicative cycles all take place at defined moments during development. In this study, we will identify the molecular regulators of endomitosis in the C. elegans intestine by performing single-cell sequencing and RNA Tomography sequencing (Tomo-seq) of intestinal cells at specific cell-cycle stages. Moreover, we will uncover the physiological importance of endomitosis and the functional difference between endomitosis and endoreplication by generating tissue-specific C. elegans mutants with alterations in the intestinal cell cycles and analyzing their intestinal function. Together, these approaches will provide an in-depth analysis of endomitosis mechanism and function, providing insights into how tissues specify cell-cycle programs and what the importance is of specific forms of polyploidization.

While cancer has been shown to rely on the accumulation of multiple mutations, conditions such as injury have been also shown to trigger cancer. Consistently, a recurrent mutation in Squamous Cell Carcinoma (SCC) in the Ras oncogene is not sufficient to trigger SCC unless either injury or a second mutation occurs. Yet, it remains elusive how wound-triggered behaviors lead to SCC. This proposal aims to identify wound-induced epithelial and niche cell behaviors that are required for cancer initiation. My post-doctoral lab has recently established an imaging approach to study how epithelial stem cell behaviors are coordinated to close a wound in live mice. I will build on this understanding and combine the study of wound repair in the context of a mouse model expressing activated-Ras mutation in epithelial stem cells. Specifically, I will determine what changes in cell behaviors, such as proliferation and migration, occur in the Ras model and I will use drug-based and genetic manipulation models to determine the functional impact of examined mutant behaviors towards SCC initiation. Furthermore, I will interrogate and manipulate the immune response, a central component of both the wound repair process and SCC, in the epithelial Ras mutant model to dissect the immune cell role during SCC insurgence. Our innovative live imaging approaches to track my populations of interest in combination with manipulative approaches will allow me to identify the relevant cells and behaviors towards cancer initiation. All together my project will provide new insights into how the physiological process of wound repair can be hijacked into a process of malignant transformation.

My goal is to understand how the cadherin-catenin complex, a force-responsive protein assembly that links neighboring cells, transduces extracellular forces into intracellular signaling events. Thus far, absolute force magnitude has been considered the main control parameter in mechanotransduction; but because mechanical cues in tissues and organs are subject to tight temporal control, it is likely that force timescales complement force magnitude to modulate junctional remodeling and downstream signal transduction. For example, during Drosophila embryogenesis regular contraction forces mediate the formation of the ventral furrow and gastrulation. In this project, I will characterize, for the first time to my knowledge, the response of intercellular adhesions to not only force amplitude, but also its temporal variation. Specifically, I will combine single-molecule force spectroscopy and single-molecule fluorescence microscopy techniques to directly quantify the kinetics of individual cadherin-catenins complexes in cells upon temporally controlled mechanical stimuli. Using this approach, I will determine how the cadherin-catenin complex integrates temporal cues into the mechanotransduction process. As a working model, I propose that cadherin clustering responds both the force magnitude and timescale, and that clustering is the key event that drives downstream signal transduction. My long-term goal as an independent investigator is to use the approaches outlined in this proposal to understand how macromolecular complexes produce, detect and respond to forces to accomplish biological processes such as mitosis, cell migration and mechanotransduction.

Life involves a myriad of inter-related chemical processes, almost none of which would proceed at adequate rates without the assistance of enzymes. Enzymes are “Nature’s catalysts”, accelerating the rates of these reactions by up to ~20 orders of magnitude, thus making life as we know it possible. No consensus has yet emerged about the ultimate origin of the "catalytic power" of enzymes, with a rather wide variety of hypotheses put forward over the years. In addition, the limitations in our understanding of enzyme catalysis are highlighted by our inability to reproduce the catalytic power of the best naturally occurring enzymes in any human made catalyst (including de novo designed enzymes); as Feynman famously pointed out, "What I cannot create, I do not understand". This lack of a definitive understanding in the origins of enzyme catalysis is most likely related to the fact that natural enzymes are the complex outcome of natural selection operating over vast expanses of time in an evolutionary process that may be determined to some extent by contingency. Here, we posit that an in-depth understanding of the catalytic power of enzymes will only be possible when we have at our disposal a procedure that routinely reproduces the emergence of new functions from non-catalytic (or minimally catalytic) scaffolds in the laboratory. Therefore, during the course of this project, we will:
1) use phylogenetic analysis to reconstruct sequences of ancestral proteins from a structure having high potential to serve as a scaffold for generating novel functionalities;
2) prepare the encoded proteins in the lab and test them not just experimentally but also computationally for the biophysical and biochemical properties that confer high evolvability on a protein scaffold;
3) synthesize very large libraries of up to ten trillion (10E13) variants using highly evolvable ancestral proteins as scaffolds;
4) use ultra-high-throughput in vitro and in vivo methodologies to screen these libraries for non-natural functions;
5) use both experimental and multiscale modeling tools to characterize the resulting de novo enzymes for the biophysical, biochemical and structural features relevant for efficient catalysis;
6) use the above characterization to test different hypotheses about the evolutionary and molecular origin of enzyme catalysis.

While there is growing realization that sleep is important for memory consolidation, most evidence remains correlative and the causal role of specific sleep events remains unclear. Moreover, there is a substantial gap between detailed electrophysiological animal research and non-invasive cognitive human research – where recording neural activity in the medial temporal lobe is particularly challenging. The proposed research aims to understand the causal role that sleep plays in memory consolidation. Upon informed consent, epilepsy patients at UCLA implanted with intracranial electrodes for clinical monitoring will participate in learning/memory paradigms combined with recordings and intracranial electrical stimulation during sleep. I will address two specific aims: (1) Determine which sleep activities correlate best with learning and memory improvements: I hypothesize that for declarative memory, NREM sleep will be best correlated with learning, and that coordinated coupling between neocortical slow waves, sleep spindles and hippocampal ripples, co-occurring with reactivation of neuronal ensembles that were selectively engaged in the learning task, will be maximally correlated with successful learning. (2) Elucidate the causal mechanisms through intracranial electrical stimulation. I hypothesize that electrical stimulation in the neocortex that mimics or locks to endogenous slow wave up-states and hippocampal ripples will be effective in promoting memory consolidation. This proposal has the potential to generate a breakthrough in our understanding of the coordination between neocortex and hippocampus during sleep and how sleep causally aids human memory consolidation.

Cell motility plays an important role in a multitude of biological processes, including embryogenesis, wound healing, and cancer metastasis. As cells move, they exert forces onto their environment and become increasingly polarized, resulting in an elongated shape and signaling and cytoskeletal components that become localized in the anterior or posterior of the cell. It is currently unclear how force generation is correlated with the distribution of key signaling components and how cells regulate their polarity. To address these questions requires a careful and precise control of external cues and the ability to link molecular events to the cell's macroscopic behavior and the generation of forces. In this project, I will quantify the establishment and reversal of cell polarity during chemotaxis using innovative microfluidic devices together with traction force microscopy (TFM) and confocal microscopy. Specifically, I will expose unpolarized Dictyostelium discoideum cells to carefully controlled chemoattractant gradients generated in microfluidic devices. I will quantify the kinetics of several key signaling components and of traction forces using confocal microscopy and TFM, respectively. Furthermore, I will examine how polarity is reversed by rapidly changing the gradient experienced by cells that migrate in specially designed 1D micropatterned tracks. By correlating the dynamics of the markers with the observed traction force patterns, I will be able to determine how cell polarity is established and reversed. Our experiments should result in a more comprehensive understanding of cell polarity, a crucial component of cell motility in many eukaryotic cell types.

The muscular systems of humans and chimpanzees differ in a variety of traits, with human muscles built for dexterity and endurance, whereas chimpanzee muscles are built for strength. However, the genetic basis of this variation is little understood. Here, I propose to use a novel system of hybrid human-chimpanzee tetraploid induced pluripotent cells to study how the regulation of genes affecting muscle physiology has evolved along our lineage. This hybrid cell line offers a unique and unprecedented opportunity to tease out genetic from environmental factors, as well as cis- from trans-acting factors. I will first differentiate the parental and hybrid cells into skeletal myocytes, the main cells determining fitness capabilities. Then, I will identify allele-specific expression by comparing the gene expression levels of the hybrids to their parental cells. This will allow me to establish the first catalogue of cis- and trans-regulatory divergence in muscle cells. Finally, I offer to pick ten genes, where it is most probable to identify the variant the drives their diverged regulation, and to genetically edit them to investigate the phenotypic effect of each sequence change. This novel system will allow me to circumnavigate a key ethical barrier in great ape studies – obtaining great ape cells. Moreover, this system and pipeline could be used to study a variety of cells, thus establishing a novel approach to identify the regulatory variants that drove species-specific traits.

While hematopoietic stem cell (HSC) transplantation has become the standard of care for treating many hematologic diseases, wider use is restrained by immunologic incompatibility and inefficient reconstitution of hematopoiesis from allogeneic sources. Pluripotent stem cells (PSCs) promise an attractive alternative source of HSCs, but current PSC-derived HSCs display fetal phenotypes that lead to stem cell exhaustion and immature progeny function. We hypothesize that human PSC-derived definitive hemogenic endothelium (HE) can be efficiently converted to HSCs with adult-like functionality. To achieve this, we first propose to develop a robust reporter system that differentially marks fetal versus adult HSCs. By combining this with recently-developed technology to transgenically convert HE into fetal HSC-like cells, we can execute in vivo targeted and high-throughput genetic screens to identify TFs and pathways that regulate HSC maturity. We will subsequently find small molecules or morphogens that modulate activity of these regulatory factors to establish transgene-free conditions for the specification of adult-like HSCs. To ensure physiological relevance of the resultant cells, they will be assessed for long-term engraftment capability, functionality of their terminally-differentiated cells and appropriate response to stress conditions in vivo. This proposal will provide proof-of-concept that adult functionality can be conferred to PSC-derived cells, provide a sustainable platform for modeling adult hematologic diseases, and most importantly, bring us one step closer towards clinical grade PSC-derived HSCs for transplantation.

Tissue stem cells (SCs) depend upon their niche microenvironment, suggesting that spatially localized signals play key roles in determining SC identity and plasticity. Evidence from the hematopoietic system suggests that vasculature is a critical niche component. Vasculature also governs tumor-initiating SCs of squamous cell carcinomas. The spatial and temporal organization of capillary networks in healthy dermis suggests that vasculature also impacts normal hair follicle SC (HFSC) behavior, a notion further fueled by hair growth sensitivity to systemic changes. Despite several recent advances, the molecular basis by which a SC can sustain regeneration is still poorly understood. Here, I aim to perform the first spatiotemporal screening to dissect the molecular crosstalk between HFSCs and vasculature remodeling, and elucidate how crosstalk dictates ‘stemness’ identity to meet regenerative demands during normal hair cycling and wound-repair. Using mouse as model, I aim to tackle these objectives by: (1) Mapping and defining the spatial and temporal localization of the skin vasculature network relative to HFSCs during development and hair cycling. (2) Elucidate the molecular crosstalk between vasculature and HFSCs, and how it changes during capillary network remodeling. (3) Dissect mechanistically how vasculature-mediated signaling influences SC physiology during homeostasis and wound-repair by using in utero lentiviral delivery to skin epithelium of HFSC-specific activation of Cas9/guide gene silencing. By understanding how the vasculature dynamically redistributes and reshapes the niche, I hope to uncover new avenues for therapeutics in skin regeneration and wound-repair.

Cell survival requires efficient adaptation to external changes. Bacteria use complex signaling networks to sense and respond to these changes, particularly nutrient availability. The current knowledge of bacterial responses to carbon starvation is based mainly on E. coli and involves small molecule second messengers like cAMP. However, intracellular responses to starvation and cAMP regulation may differ in other species. The alpha-proteobacterium Caulobacter crescentus is a powerful model for understanding how intracellular signals are transduced and impact cell-cycle progression in bacteria. Carbon starvation somehow prevents cell differentiation and a G1-S transition during the Caulobacter cell cycle. This differentiation step is normally triggered by the synthesis of a regulator called SpmX. I propose that the absence of a carbon source prevents SpmX synthesis through the inhibition of its transcriptional regulator, TacA. I will use ChIP-qPCR and in vivo phosphorylation assays to test this model and dissect the mechanisms that temporally regulate SpmX. Further, I hypothesize that, like E. coli, the main intracellular signal following carbon starvation in Caulobacter is cAMP, and that cAMP directly regulates TacA activation. I will measure cAMP intracellular levels using a FRET-based method and assess the impact of cAMP on TacA activity and SpmX synthesis. I will also use capture-compound mass spectrometry to identify the cAMP interactome in Caulobacter. I anticipate that this study will provide new insights into cAMP-dependent cellular responses and nutrient-sensing mechanisms in Caulobacter, which may differ significantly from what is known in E. coli.

During development, embryonic tissues develop into organs of stereotyped size and shape. Morphogens are secreted from discrete regions in developing organs and form spatial concentration gradients that guide gene activation, pattern formation and tissue growth. Morphogen gradients scale with tissue size, ensuring that morphological patterns remain proportionate in organs of different size. The mechanisms of morphogen dispersal and scaling are not yet clear. Two models for morphogen spreading attracted much attention in the field: i) fast extracellular diffusion versus ii) cycles of internalization and recycling (transcytosis). I will develop a generalized theory of transport that captures extracellular diffusion and transcytosis simultaneously: I will consider extracellular diffusion, morphogen/receptor binding/unbinding and various endocytic trafficking steps. This novel theoretical method will uncover the parameter space where either extracellular diffusion or transcytosis dominates the dynamics of dispersal. I will also explore whether both phenomena could contribute significantly to the length scale of the steady-state gradient. My model system will be the Decapentaplegic (Dpp) gradient in the fly wing. Using novel experimental endocytic assays from the host lab, I will parameterize Dpp trafficking dynamics to establish the prevalence of extracellular diffusion versus transcytosis. Unpublished work from the host lab shows that trafficking is spatially heterogeneous in the wing. I will upgrade my theory to a heterogeneous scenario with the potential to yield, as an emerging property, a novel mechanism of gradient scaling.

When we navigate and make decisions in familiar environments we rely on an internal, mental map of the features and general shape of the environment. Remarkably, work over the last four decades has revealed how different locations in space and different navigational plans are represented as patterns of activity in thousands of neurons in the brains of rodents. It stands to reason that this 'neural representation' of an environment should remain stable as long as an animal is using it, to allow the animal the reliably navigate. However, we found that the neural activity that makes up this map changes over time so that different neurons represent specific features of a familiar environment after several days. We do not have a theory to explain how the brain can build and use such a continually changing mental map, nor do we know why the map changes, or what purpose this continual change might serve. In this project we will build and experimentally test new theories and models of how the brain represents environments, potentially changing our understanding of how the brain works.

The random selection of an allele to be exclusively expressed affects gene expression and any possible phenotypes related to the allele. This may be one of the key mechanisms through which phenotypic variation is established. Estimates of the prevalence of autosomal random monoallelic expression (aRME) vary widely between different experiments and cell types and importantly, the frequencies of fixed and dynamic aRME have not been addressed in the same experiments. I aim to address the balance between fixed and dynamic aRME in an in vivo as well as in vitro model. By combining single cell, allele sensitive transcriptomics approaches, which have been developed in the Sandberg lab, combined with an inducible cell-lineage reporter system, I will be uniquely able to characterize the extent of both fixed and dynamic aRME. We hypothesize that the balance between fixed and dynamic aRME depends on the stabilization of expression on a given allele. According to that model, fixed as well as dynamic aRME is caused by stochastic expression that is then stabilized to different extents. By studying this in vivo we will be able to characterize the stability of aRME in different tissues where aRME stabilization may differ (eg. neurons vs. fibroblasts). The in vitro model on the other hand would give me the possibility to perturb the system and study the mechanisms that establish and stabilize aRME. Additionally, the data resulting from the experiments I propose will also provide a novel and unique insight into the relationship between gene expression and cell clonality in a wide range of mouse tissues and cell types.

Chemical probes are small molecules that can perturb the activity of a target macromolecule, and thus can be used to understand the biological roles of receptors. Currently, identifying a chemical probe for one specific target requires screening many chemicals to find ones that bind that target. If one screen could be used to find chemical probes for many targets, one could dramatically increase the rate of probe discovery.

Our first goal is to identify novel chemical probes for approximately 40 receptors in S. cerevisiae. This will be done using a novel high throughput metabolomics approach. Recent advances in the Sauer lab allow for the speedy generation of a metabolomic fingerprint of the effect of a compound treatment on the cell. Therefore it should be possible to adopt an innovative approach to chemical screening where compounds are identified that are able to mimic the metabolomic profile of yeast where receptors are genetically perturbed. This will allow us to find chemical probes for those targets. Thus, we will be able to identify many chemical probes for many receptors from a single chemical screen.

Our second goal will be to investigate the roles of these receptors in vivo using our chemical probes. We will investigate what specific metabolomic and transcriptomic changes individual S. cerevisiae nutrient sensors elicit when they activated by a non-metabolizable ligand. This is of interest because it can isolate what the cellular response to receptor activation is, and potentially show why it is adaptive for the cell to respond in that way. Additionally, we will use them to investigate the biological roles of poorly characterized receptors in S. cerevisiae.

The neural circuits underlying vision share a lot of structural and functional similarities between Drosophila and mammals, with motion detection being the best-studied visual modality. The highly stereotyped columnar and layered organization of Drosophila optic lobe neuropils (the visual processing centres) allows a precise study of circuit wiring. Two parallel motion channels have been described in Drosophila, one for detecting moving light edges (ON pathway) and one for detecting moving dark edges (OFF pathway). The main components of these channels have been identified, comprising neurons in the lamina, medulla, lobula and lobula plate neuropils. However, little is known about the specific mechanisms leading to the orderly and stereotypic assembly of these cell types to form functional circuits. In this proposal we offer to study by live imaging the spatiotemporal dynamics of motion detection circuit assembly and the determinants of such formation. It has been proposed that circuits are assembled in the order of neuron generation. To test this hypothesis, we will analyse the assembly order of ON and OFF motion circuit components and study how the developmental specification of cell types is linked to circuit formation. Furthermore, we will investigate whether there are neurons with pioneer roles in circuit assembly, the influence of neuronal soma positioning on axonal and dendritic layer targeting, and the role of spontaneous activity in circuit wiring. Altogether, this project will advance our understanding of the mechanisms governing functional circuitry assembly, which is likely to have broader implications for other neuronal systems.

Autophagy is a ubiquitous process of eukaryotic cell biology, which occurs by the formation and growth of the isolation membrane (IM). The IM engulfs bulk cytosol or targets such as protein aggregates, damaged organelles, and invading pathogens. The IM closes to become the autophagosome, fuses with the lysosome, and its contents are degraded. More than 40 Atg proteins are dedicated to autophagy, thus the “parts list” is known. How the autophagosome initiates, grows, and closes, using these parts is a world-class mystery. We propose to use biochemical reconstitution, cell imaging, and computational biophysics to dissect the uptake of the intracellular pathogen Salmonella (“xenophagy”) and so reveal the physical basis of autophagosome formation. In one major model, derived from yeast data, the IM nucleates from several small vesicles that then fuse into a sheet. The sheet grows by fusion of more small vesicles. In another model, derived from mammalian cell data, the IM is extruded out of a domain of the endoplasmic reticulum (ER). We propose to resolve which of these mechanisms is operative, by reconstituting autophagosome formation, and by directly imaging the process both in vitro and in cells using real-time super-resolution imaging. We will then move beyond distinguishing between these qualitative models, to a quantitative, biophysical paradigm. The arrival, departure, copy numbers, and structures of the relevant protein complexes on and off membranes will be determined in vitro and in vivo. Changes in membrane shape will be correlated with these parameters. These will be input to a computational model accounting for membrane physical properties and protein structures. The complexity of this system, with 10s-100s of copies of more than 40 different proteins, operating on a time scale of minutes and a length scale of two microns, makes this one of the most complex systems to be tackled with such biochemical and biophysical detail.

This project will investigate locomotion control in undulatory animals and transitions between different environmental media. It will test the idea that a single control principle could explain different modes of locomotion in vertebrate and invertebrate animals for different morphologies in different environmental media (water and ground). The control principle is based on “sensory synchronization” of local neural oscillators. It merges ideas from central pattern generators and from sensory-driven locomotor models. We postulate that flexible motor patterns for body-limb coordination during swimming, crawling and walking can be obtained in a distributed network of neural oscillators that rely to a large extent on (multimodal) sensory feedback signals for synchronization. In other words, we postulate that the synchronization between oscillations in body and limbs, and hence the generation of gaits, is due to mechanical interactions of the body with the environment and the resulting sensory signals more than to direct neuronal (or central) coupling between the involved neuronal oscillators. We also postulate that many observed features of various gaits are due to changes in the environment rather than to changes in the descending commands.
This idea and related predictions will be tested by investigating Polypterus (a walking fish), salamander, and centipede locomotion. These three species represent a diversity of life that has overcome the problems associated with neuro-control in amphibious environments. Our research tools combine animal locomotion studies, neuromechanical simulations, and robotic experiments. The neuromechanical simulations and the robots will play an essential role in decoding the complex interplay between central pattern generators, sensory feedback, the musculoskeletal system, and the different environments. The interdisciplinary approach will allow us to investigate whether the same control principle could not only facilitate adaptations to growth, lesions, perturbations, and different environments during the lifetime of the animal, but also to evolutionary adaptations (e.g. changes of ecological niches and of morphologies) in particular the transition from aquatic to terrestrial environments in vertebrates and invertebrates, a key step in evolution.

In photosynthesis, light-harvesting complexes (LHCs) capture solar energy and feed it to the downstream molecular machinery. However, when light absorption exceeds the capacity for utilization, the excess energy can cause damage. Thus, LHCs have evolved a feedback loop that triggers photoprotective energy dissipation. The critical importance of photoprotection for plant fitness has been demonstrated, as well as its impact on crop yields. However, the mechanisms of photoprotection ? from fast chemical reactions of molecules to slow conformational changes of proteins ? have not yet been resolved. Despite extensive studies of the native photosynthetic apparatus, previous experiments have not been able to control architecture in order to distinguish between hypotheses of the mechanisms of photoprotection. To overcome this barrier, we take a novel synthetic structural biology approach. We build minimal regulatory units using model membranes that we measure and model to gain a multi-timescale understanding of photoprotection, which is core to natural light harvesting. Taking this synthetic structural biology approach is a significant experimental challenge in three distinct fields; biochemical and molecular biology development is required to construct a well-defined and well-characterized sample, new spectroscopic analyses and collection strategies are required to adapt optical experiments to probe these samples, and theoretical advances are required to model these length scales.
Uniquely, photoprotection regulates a photophysical process, providing an intrinsic spectroscopic handle to uncover how information flows within the multi-timescale phenomena of biological regulation. LHCs and other biological systems perform meso- and macroscopic functions efficiently and robustly even when exposed to the changing and fluctuating environment that is a hallmark of natural systems. Furthermore, biological systems adapt and even reprogram their function in response to the environment. Unveiling the design and working principles of such responsive and highly autonomous behaviors exhibited by molecular systems is one of today’s grand challenge areas from the standpoint of both theoretical and experimental research; we incisively explore such behaviors in the context of photosynthetic light harvesting through the work proposed here.

Mitochondria are organelles that perform fundamental functions in eukaryotic cells. They are responsible of cellular respiration, energy generation and an array of biosynthetic processes. Additionally, they play a key regulatory role in programmed cell death. In medicine they are widely studied, as mitochondrial dysfunction/disease have enormous health consequences, due to their involvement in aging and in a number of pathologies including, cancer, metabolic disorders and heart disease.
Despite their importance, and the effort of numerous researches, multiple aspects of mitochondrial biology are still unclear. In this project we propose to exploit a unique system to develop novel tools to study and to interact with mitochondria, with potential groundbreaking applications. Midichloria mitochondrii is an intracellular bacterium that presents a characteristic that makes it unique: it is able to enter the mitochondria of its host cells. This unique interaction has been discovered recently, and has still widely uncharacterized. Our multidisciplinary team envisioned an integrated experimental approach, using state of the art technologies in the fields of bioinformatics, -omics, imaging, biochemistry, and cellular biology to investigate this novel biological system.
This research project will allow us to understand the evolution and the functional details of the relationship between M. mitochondrii and its host at the organismal, cellular and molecular level. We will use this novel information to develop an in vitro model system that will allow us to use this unique bacterium to interact with mitochondria in novel ways. This approach will ultimately lead to the development of new experimental tools to characterize mitochondrial mechanisms, including mitochondrial fusion/fission, inner and outer membrane repair, metabolism and apoptotic signaling.

Demography (i.e. a population’s history of size changes, gene flow, and divergence events) affects patterns of genetic diversity, and has implications for studying evolution, adaptations, and disease. However, demographic inference is complicated by “ghost populations,” unobserved populations in the past that admixed with the ancestors of sampled populations. Inferring this hidden structure is challenging both because the number of ghost populations is unknown, and also because signals of cryptic structure may be distorted by other evolutionary forces. I will address this problem by developing a flexible method capable of estimating unknown population structure jointly with other sources of genetic diversity. Our method will combine models from Coalescent theory with machine learning techniques for inferring graph structure, and will estimate the number of ghost populations, their demographic history, and the distribution of fitness effects for mutations arising in these ghost populations. Furthermore, our approach will allow us to estimate allele frequencies in ancestral populations, which will be useful for identifying adaptive introgressed regions from ghost populations, and for addressing questions such as the adaptive consequences of gene flow from ghost populations. We will apply our new method to diverse human whole-genomes to shed light on open questions in recent human evolutionary history, such as the relationship between modern African populations and the out-of-Africa population that gave rise to modern Europeans and Asians, or the extent to which variants selected in particular populations have spread more broadly by admixture and introgression.