Explore the vast range of titles available.

Subclonal reconstruction algorithms use bulk DNA sequencing data to quantify parameters of tumor evolution, allowing an assessment of how cancers initiate, progress and respond to selective pressures. We launched the ICGC–TCGA (International Cancer Genome Consortium–The Cancer Genome Atlas) DREAM Somatic Mutation Calling Tumor Heterogeneity and Evolution Challenge to benchmark existing subclonal reconstruction algorithms. This 7-year community effort used cloud computing to benchmark 31 subclonal reconstruction algorithms on 51 simulated tumors. Algorithms were scored on seven independent tasks, leading to 12,061 total runs. Algorithm choice influenced performance substantially more than tumor features but purity-adjusted read depth, copy-number state and read mappability were associated with the performance of most algorithms on most tasks. No single algorithm was a top performer for all seven tasks and existing ensemble strategies were unable to outperform the best individual methods, highlighting a key research need. All containerized methods, evaluation code and datasets are available to support further assessment of the determinants of subclonal reconstruction accuracy and development of improved methods to understand tumor evolution. Benchmarking of tumor subclonal reconstruction algorithms finds that interalgorithm variability overwhelms intertumor variability and identifies key quality metrics.

According to this traditional model, the selective advantage is conferred by a small set of driver mutations, but as the subclones that bear them successively expand, they also accumulate passenger mutations, which can be detected in sequencing experiments1. Genomes of individual tumors contain hundreds to many thousands of these genetic variants at a wide range of frequencies5,6. Because genetic drift can drive novel variants to high frequencies, it is of great interest to discern the relative importance of selection and drift in shaping the frequency distribution of variants in any given tumor. [...] the deterministic model of tumor growth described by Williams et al. relies on strong biological assumptions, including synchronous cell divisions, constant cell death, and constant mutation and division rates. [...]discrimination of neutral and non-neutral simulated tumors by using a linear fit is almost arbitrary, with 53.5% false-positive neutral calls in nonneutral tumors (Fig. 1b) and an area under the receiver operating characteristic curve of 0.42 for the classification of 1,919 neutral and 1,919 non-neutral tumors (Fig. 1c).

Time series data play a crucial role in many applications, such as biology, medicine, economics, engineering and others. One of the most powerful approaches for time series is latent-variable models, thanks to their ability to handle multi-dimensional data with complex interactions. Typically, these models represent a timeline as a sequence of discrete states and therefore assume that observations occur at regular intervals. However, this assumption does not always hold. An illustrative example is medical records, where a patient is screened only when the need arises, resulting in the irregularly-spaced and possibly sparse time series. In this type of time series, the time intervals between the observations can provide valuable information about the time series, such as patient's health condition. To bridge the gap between the data and the available models, a common approach is to convert a continuous timeline into a discrete one by aggregating observations into a sequence of discrete clusters. However, this transformation erases a lot of the temporal structure of the data and prevents us from utilizing the data to its full potential. In this thesis, I present latent-variable models for continuous-time data across several application domains. First, I present a method for ordering cancer mutations on a linear timeline and use a mixture model to summarize them into a set of trajectories over time. Thanks to this ordering, I perform a more fine-grained discretization of the cancer timeline in comparison to the previous methods and can more accurately detect the changes in cancer dynamics. Next, I introduce Latent Ordinary Differential Equations (Latent ODE) -- a framework that allows to model time series as a solution of a differential equation, in other words, as a continuous function over time. Unlike the previous models, this approach does not require any discretization of the data and can naturally handle irregularly-spaced time points. I showcase the potential of the model on the interpretable dataset. I demonstrate that the Latent ODE model has better extrapolation properties and is more robust to noise compared to existing sequential models. Finally, I improve the Latent ODE model by proposing an ODE-based recognition model. I demonstrate that, by preserving the observations on a real-values timeline and modelling them as a continuous function, we can get improvement in a variety of tasks, such as forecasting, imputation and classification.

Intra-tumor heterogeneity (ITH) is a mechanism of therapeutic resistance and therefore an important clinical challenge. However, the extent, origin and drivers of ITH across cancer types are poorly understood. To address this question, we extensively characterize ITH across whole-genome sequences of 2,658 cancer samples, spanning 38 cancer types. Nearly all informative samples (95.1%) contain evidence of distinct subclonal expansions, with frequent branching relationships between subclones. We observe positive selection of subclonal driver mutations across most cancer types, and identify cancer type specific subclonal patterns of driver gene mutations, fusions, structural variants and copy-number alterations, as well as dynamic changes in mutational processes between subclonal expansions. Our results underline the importance of ITH and its drivers in tumor evolution, and provide an unprecedented pan-cancer resource of comprehensively annotated subclonal events from whole-genome sequencing data. Competing Interest Statement R.B. owns equity in Ampressa Therapeutics. G.G. receives research funds from IBM and Pharmacyclics and is an inventor on patent applications related to MuTect, ABSOLUTE, MutSig, MSMuTect and POLYSOLVER. I.L. is a consultant for PACT Pharma. B.J.R. is a consultant at and has ownership interest (including stock and patents) in Medley Genomics. All other authors declare no competing interests.

Amazing progress has been made in controllable generative modeling, especially over the last few years. However, some challenges remain. One of them is precise and iterative object editing. In many of the current methods, trying to edit the generated image (for example, changing the color of a particular object in the scene or changing the background while keeping other elements unchanged) by changing the conditioning signals often leads to unintended global changes in the scene. In this work, we take the first steps to address the above challenges. Taking inspiration from the Universal Scene Descriptor (USD) standard developed in the computer graphics community, we introduce the \"Neural Universal Scene Descriptor\" or Neural USD. In this framework, we represent scenes and objects in a structured, hierarchical manner. This accommodates diverse signals, minimizes model-specific constraints, and enables per-object control over appearance, geometry, and pose. We further apply a fine-tuning approach which ensures that the above control signals are disentangled from one another. We evaluate several design considerations for our framework, demonstrating how Neural USD enables iterative and incremental workflows. More information at: https://escontrela.me/neural_usd .

Simulating large scenes with many rigid objects is crucial for a variety of applications, such as robotics, engineering, film and video games. Rigid interactions are notoriously hard to model: small changes to the initial state or the simulation parameters can lead to large changes in the final state. Recently, learned simulators based on graph networks (GNNs) were developed as an alternative to hand-designed simulators like MuJoCo and PyBullet. They are able to accurately capture dynamics of real objects directly from real-world observations. However, current state-of-the-art learned simulators operate on meshes and scale poorly to scenes with many objects or detailed shapes. Here we present SDF-Sim, the first learned rigid-body simulator designed for scale. We use learned signed-distance functions (SDFs) to represent the object shapes and to speed up distance computation. We design the simulator to leverage SDFs and avoid the fundamental bottleneck of the previous simulators associated with collision detection. For the first time in literature, we demonstrate that we can scale the GNN-based simulators to scenes with hundreds of objects and up to 1.1 million nodes, where mesh-based approaches run out of memory. Finally, we show that SDF-Sim can be applied to real world scenes by extracting SDFs from multi-view images.

Accurately simulating real world object dynamics is essential for various applications such as robotics, engineering, graphics, and design. To better capture complex real dynamics such as contact and friction, learned simulators based on graph networks have recently shown great promise. However, applying these learned simulators to real scenes comes with two major challenges: first, scaling learned simulators to handle the complexity of real world scenes which can involve hundreds of objects each with complicated 3D shapes, and second, handling inputs from perception rather than 3D state information. Here we introduce a method which substantially reduces the memory required to run graph-based learned simulators. Based on this memory-efficient simulation model, we then present a perceptual interface in the form of editable NeRFs which can convert real-world scenes into a structured representation that can be processed by graph network simulator. We show that our method uses substantially less memory than previous graph-based simulators while retaining their accuracy, and that the simulators learned in synthetic environments can be applied to real world scenes captured from multiple camera angles. This paves the way for expanding the application of learned simulators to settings where only perceptual information is available at inference time.

In the area of physical simulations, nearly all neural-network-based methods directly predict future states from the input states. However, many traditional simulation engines instead model the constraints of the system and select the state which satisfies them. Here we present a framework for constraint-based learned simulation, where a scalar constraint function is implemented as a graph neural network, and future predictions are computed by solving the optimization problem defined by the learned constraint. Our model achieves comparable or better accuracy to top learned simulators on a variety of challenging physical domains, and offers several unique advantages. We can improve the simulation accuracy on a larger system by applying more solver iterations at test time. We also can incorporate novel hand-designed constraints at test time and simulate new dynamics which were not present in the training data. Our constraint-based framework shows how key techniques from traditional simulation and numerical methods can be leveraged as inductive biases in machine learning simulators.

A current limitation of video generative video models is that they generate plausible looking frames, but poor motion -- an issue that is not well captured by FVD and other popular methods for evaluating generated videos. Here we go beyond FVD by developing a metric which better measures plausible object interactions and motion. Our novel approach is based on auto-encoding point tracks and yields motion features that can be used to not only compare distributions of videos (as few as one generated and one ground truth, or as many as two datasets), but also for evaluating motion of single videos. We show that using point tracks instead of pixel reconstruction or action recognition features results in a metric which is markedly more sensitive to temporal distortions in synthetic data, and can predict human evaluations of temporal consistency and realism in generated videos obtained from open-source models better than a wide range of alternatives. We also show that by using a point track representation, we can spatiotemporally localize generative video inconsistencies, providing extra interpretability of generated video errors relative to prior work. An overview of the results and link to the code can be found on the project page: http://trajan-paper.github.io.

Realistic simulation is critical for applications ranging from robotics to animation. Traditional analytic simulators sometimes struggle to capture sufficiently realistic simulation which can lead to problems including the well known \"sim-to-real\" gap in robotics. Learned simulators have emerged as an alternative for better capturing real-world physical dynamics, but require access to privileged ground truth physics information such as precise object geometry or particle tracks. Here we propose a method for learning simulators directly from observations. Visual Particle Dynamics (VPD) jointly learns a latent particle-based representation of 3D scenes, a neural simulator of the latent particle dynamics, and a renderer that can produce images of the scene from arbitrary views. VPD learns end to end from posed RGB-D videos and does not require access to privileged information. Unlike existing 2D video prediction models, we show that VPD's 3D structure enables scene editing and long-term predictions. These results pave the way for downstream applications ranging from video editing to robotic planning.