Offline multi-agent reinforcement learning (MARL) is severely hampered by the challenge of evaluating out-of-distribution (OOD) joint actions. Our core finding is that when the behavior policy is factorized—a common scenario where agents act fully or partially independently during data collection—a strategy of partial action replacement (PAR) can significantly mitigate this challenge. PAR updates a single or part of agents' actions while the others remain fixed to the behavioral data, reducing distribution shift compared to full joint-action updates. Based on this insight, we develop Soft-Partial Conservative Q-Learning (SPaCQL), using PAR to mitigate OOD issue and dynamically weighting different PAR strategies based on the uncertainty of value estimation. We provide a rigorous theoretical foundation for this approach, proving that under factorized behavior policies, the induced distribution shift scales linearly with the number of deviating agents rather than exponentially with the joint-action space. This yields a provably tighter value error bound for this important class of offline MARL problems. Our theoretical results also indicate that SPaCQL adaptively addresses distribution shift using uncertainty-informed weights. Our empirical results demonstrate SPaCQL enables more effective policy learning, and manifest its remarkable superiority over baseline algorithms when the offline dataset exhibits the independence structure.

This paper explores the impact of relational state abstraction on sample efficiency and performance in collaborative Multi-Agent Reinforcement Learning. The proposed abstraction is based on spatial relationships in environments where direct communication between agents is not allowed, leveraging the ubiquity of spatial reasoning in real-world multi-agent scenarios. We introduce MARC (Multi-Agent Relational Critic), a simple yet effective critic architecture incorporating spatial relational inductive biases by transforming the state into a spatial graph and processing it through a relational graph neural network. The performance of MARC is evaluated across four collaborative tasks, including a novel environment with heterogeneous agents. We conduct a comprehensive empirical analysis, comparing MARC against state-of-the-art MARL baselines, demonstrating improvements in both sample efficiency and asymptotic performance, as well as its potential for generalization. Our findings suggest that a minimal integration of spatial relational inductive biases as abstraction can yield substantial benefits without requiring complex designs or task-specific engineering. This work provides insights into the potential of relational state abstraction to address sample efficiency, a key challenge in MARL, offering a promising direction for developing more efficient algorithms in spatially complex environments.

Offline reinforcement learning (RL) seeks to learn optimal policies from static datasets without interacting with the environment. A common challenge is handling multi-modal action distributions, where multiple behaviours are represented in the data. Existing methods often assume unimodal behaviour policies, leading to suboptimal performance when this assumption is violated. We propose weighted imitation Learning on One Mode (LOM), a novel approach that focuses on learning from a single, promising mode of the behaviour policy. By using a Gaussian mixture model to identify modes and selecting the best mode based on expected returns, LOM avoids the pitfalls of averaging over conflicting actions. Theoretically, we show that LOM improves performance while maintaining simplicity in policy learning. Empirically, LOM outperforms existing methods on standard D4RL benchmarks and demonstrates its effectiveness in complex, multi-modal scenarios.

Growing data privacy demands, driven by regulations like GDPR and CCPA, require machine unlearning methods capable of swiftly removing the influence of specific training points. Although verified approaches like SISA, using data slicing and checkpointing, achieve efficient unlearning for single models by reverting to intermediate states, these methods struggle in teacher-student knowledge distillation settings. Unlearning in the teacher typically forces costly, complete student retraining due to pervasive information propagation during distillation. Our primary contribution is PURGE (Partitioned Unlearning with Retraining Guarantee for Ensembles), a novel framework integrating verified unlearning with distillation. We introduce constituent mapping and an incremental multi-teacher strategy that partitions the distillation process, confines each teacher constituent’s impact to distinct student data subsets, and crucially maintains data isolation. The PURGE framework substantially reduces retraining overhead—requiring only partial student updates—when teacher-side unlearning occurs. We provide both theoretical analysis, quantifying significant speed-ups in the unlearning process, and empirical validation on multiple datasets, demonstrating that PURGE achieves these efficiency gains while maintaining student accuracy comparable to standard baselines.

Retrosynthesis planning aims to decompose target molecules into available building blocks, forming a synthetic tree where each internal node represents an intermediate compound and each leaf ideally corresponds to a purchasable reactant. However, this tree becomes invalid if any leaf node is not a valid building block, making the planning process vulnerable to the "weakest link" in the synthetic route. Existing methods often optimise for average performance across branches, failing to account for this worst-case sensitivity. In this paper, we reframe retrosynthesis as a worst-path optimisation problem within tree-structured Markov Decision Processes (MDPs). We prove that this formulation admits a unique optimal solution and provides monotonic improvement guarantees. Building on this insight, we introduce Interactive Retrosynthesis Planning (InterRetro), a method that interacts with the tree MDP, learns a value function for worst-path outcomes, and improves its policy through self-imitation, preferentially reinforcing past decisions with high estimated advantage. Empirically, InterRetro achieves state-of-the-art results — solving 100\% of targets on the Retro*-190 benchmark, shortening synthetic routes by 4.9\%, and achieving promising performance using only 10\% of the training data.

In reinforcement learning with sparse rewards, demonstrations can accelerate learning, but determining when to imitate them remains challenging. We propose Smooth Policy Regularisation from Demonstrations (SPReD), a framework that addresses the fundamental question: when should an agent imitate a demonstration versus follow its own policy? SPReD uses ensemble methods to explicitly model Q-value distributions for both demonstration and policy actions, quantifying uncertainty for comparisons. We develop two complementary uncertainty-aware methods: a probabilistic approach estimating the likelihood of demonstration superiority, and an advantage-based approach scaling imitation by statistical significance. Unlike prevailing methods (e.g. Q-filter) that make binary imitation decisions, SPReD applies continuous, uncertainty-proportional regularisation weights, reducing gradient variance during training. Despite its computational simplicity, SPReD achieves remarkable gains in experiments across eight robotics tasks, outperforming existing approaches by up to a factor of 14 in complex tasks while maintaining robustness to demonstration quality and quantity. Our code is available at https://github.com/YujieZhu7/SPReD.

Abstract In human society, the conflict between self-interest and collective well-being often obstructs efforts to achieve shared welfare. Related concepts like the Tragedy of the Commons and Social Dilemmas frequently manifest in our daily lives. As artificial agents increasingly serve as autonomous proxies for humans, we propose a novel multi-agent reinforcement learning (MARL) method to address this issue - learning policies to maximise collective returns even when individual agents’ interests conflict with the collective one. Unlike traditional cooperative MARL solutions that involve sharing rewards, values, and policies or designing intrinsic rewards to encourage agents to learn collectively optimal policies, we propose a novel MARL approach where agents exchange action suggestions. Our method reveals less private information compared to sharing rewards, values, or policies, while enabling effective cooperation without the need to design intrinsic rewards. Our algorithm is supported by our theoretical analysis that establishes a bound on the discrepancy between collective and individual objectives, demonstrating how sharing suggestions can align agents’ behaviours with the collective objective. Experimental results demonstrate that our algorithm performs competitively with baselines that rely on value or policy sharing or intrinsic rewards.

Precise understanding of free-text radiology reports through localised extraction of clinical findings can enhance medical imaging applications like computer-aided diagnosis. We present a new task, that of segmenting radiology reports into topically meaningful passages (segments) and a transformer-based model that both segments reports into semantically coherent segments and classifies each segment using a set of 37 radiological abnormalities, thus enabling fine-grained analysis. This contrasts with prior work that performs classification on full reports without localisation. Trained on over 2.7 million unlabelled chest X-ray reports and over 28k segmented and labelled reports, our model achieves state-of-the-art performance on report segmentation (0.0442 WinDiff) and multi-label classification (0.84 report-level macro F1) over 37 radiological labels and 8 NLP-specific labels. This work establishes new benchmarks for fine-grained understanding of free-text radiology reports, with precise localisation of semantics unlocking new opportunities to improve computer vision model training and clinical decision support. We open-source our annotation tool, model code and pretrained weights to encourage future research.

Discrete-action reinforcement learning algorithms often falter in tasks with high-dimensional discrete action spaces due to the vast number of possible actions. A recent advancement leverages value-decomposition, a concept from multi-agent reinforcement learning, to tackle this challenge. This study delves deep into the effects of this value-decomposition, revealing that whilst it curtails the over-estimation bias inherent to Q-learning algorithms, it amplifies target variance. To counteract this, we present an ensemble of critics to mitigate target variance. Moreover, we introduce a regularisation loss that helps to mitigate the effects that exploratory actions in one dimension can have on the value of optimal actions in other dimensions. Our novel algorithm, REValueD, tested on discretised versions of the DeepMind Control Suite tasks, showcases superior performance, especially in the challenging humanoid and dog tasks. We further dissect the factors influencing REValueD's performance, evaluating the significance of the regularisation loss and the scalability of REValueD with increasing sub-actions per dimension.

Abstract Offline reinforcement learning agents seek optimal policies from fixed data sets. With environmental interaction prohibited, agents face significant challenges in preventing errors in value estimates from compounding and subsequently causing the learning process to collapse. Uncertainty estimation using ensembles compensates for this by penalising high-variance value estimates, allowing agents to learn robust policies based on data-driven actions. However, the requirement for large ensembles to facilitate sufficient penalisation results in significant computational overhead. In this work, we examine the role of policy constraints as a mechanism for regulating uncertainty, and the corresponding balance between level of constraint and ensemble size. By incorporating behavioural cloning into policy updates, we show empirically that sufficient penalisation can be achieved with a much smaller ensemble size, substantially reducing computational demand while retaining state-of-the-art performance on benchmarking tasks. Furthermore, we show how such an approach can facilitate stable online fine tuning, allowing for continued policy improvement while avoiding severe performance drops.

Abstract Behavioural cloning (BC) is a commonly used imitation learning method to infer a sequential decision-making policy from expert demonstrations. However, when the quality of the data is not optimal, the resulting behavioural policy also performs sub-optimally once deployed. Recently, there has been a surge in offline reinforcement learning methods that hold the promise to extract high-quality policies from sub-optimal historical data. A common approach is to perform regularisation during training, encouraging updates during policy evaluation and/or policy improvement to stay close to the underlying data. In this work, we investigate whether an offline approach to improving the quality of the existing data can lead to improved behavioural policies without any changes in the BC algorithm. The proposed data improvement approach - Model-Based Trajectory Stitching (MBTS) - generates new trajectories (sequences of states and actions) by ‘stitching’ pairs of states that were disconnected in the original data and generating their connecting new action. By construction, these new transitions are guaranteed to be highly plausible according to probabilistic models of the environment, and to improve a state-value function. We demonstrate that the iterative process of replacing old trajectories with new ones incrementally improves the underlying behavioural policy. Extensive experimental results show that significant performance gains can be achieved using MBTS over BC policies extracted from the original data. Furthermore, using the D4RL benchmarking suite, we demonstrate that state-of-the-art results are obtained by combining MBTS with two existing offline learning methodologies reliant on BC, model-based offline planning (MBOP) and policy constraint (TD3+BC).

Abstract Offline reinforcement learning (RL) aims to create policies for sequential decision-making using exclusively offline datasets. This presents a significant challenge, especially when attempting to accomplish multiple distinct goals or outcomes within a given scenario while receiving sparse rewards. Prior methods using advantage weighting for offline goal-conditioned learning improve policies monotonically. However, they still face challenges from distribution shift and multi-modality that arise due to conflicting ways to reach a goal. This issue is especially challenging in long-horizon tasks, where the presence of multiple, often conflicting, solutions makes it hard to identify a single optimal policy for transitioning from a state to a desired goal. To address these challenges, we introduce a complementary advantage-based weighting scheme that incorporates an additional source of inductive bias. Given a value-based partitioning of the state space, the contribution of actions expected to lead to target regions that are easier to reach, compared to the final goal, is further increased. Our proposed approach, Dual-Advantage Weighted Offline Goal-conditioned RL, outperforms several competing offline algorithms in widely used benchmarks. Furthermore, we provide a theoretical guarantee that the learned policy will not be inferior to the underlying behavior policy.

Abstract In artificial multi-agent systems, the ability to learn collaborative policies is predicated upon the agents’ communication skills: they must be able to encode the information received from the environment and learn how to share it with other agents as required by the task at hand. We present a deep reinforcement learning approach, Connectivity Driven Communication (CDC), that facilitates the emergence of multi-agent collaborative behaviour only through experience. The agents are modelled as nodes of a weighted graph whose state-dependent edges encode pair-wise messages that can be exchanged. We introduce a graph-dependent attention mechanisms that controls how the agents’ incoming messages are weighted. This mechanism takes into full account the current state of the system as represented by the graph, and builds upon a diffusion process that captures how the information flows on the graph. The graph topology is not assumed to be known a priori, but depends dynamically on the agents’ observations, and is learnt concurrently with the attention mechanism and policy in an end-to-end fashion. Our empirical results show that CDC is able to learn effective collaborative policies and can over-perform competing learning algorithms on cooperative navigation tasks.

Recent years have witnessed the prevalence of memory-based methods for Semi-supervised Video Object Segmentation (SVOS) which utilise past frames efficiently for label propagation. When conducting feature matching, fine-grained multi-scale feature matching has typically been performed using all query points, which inevitably results in redundant computations and thus makes the fusion of multi-scale results ineffective. In this paper, we develop a new Point-based Memory Network, termed as PMNet, to perform fine-grained feature matching on hard samples only, assuming that easy samples can already obtain satisfactory matching results without the need for complicated multi-scale feature matching. Our approach first generates an uncertainty map from the initial decoding outputs. Next, the fine-grained features at uncertain locations are sampled to match the memory features on the same scale. Finally, the matching results are further decoded to provide a refined output. The point-based scheme works with the coarsest feature matching in a complementary and efficient manner. Furthermore, we propose an approach to adaptively perform global or regional matching based on the motion history of memory points, making our method more robust against ambiguous backgrounds. Experimental results on several benchmark datasets demonstrate the superiority of our proposed method over state-of-the-art methods.

Referring Video Object Segmentation (RVOS) aims to segment the text-depicted object from video sequences. With excellent capabilities in long-range modelling and information interaction, transformers have been increasingly applied in existing RVOS architectures. To better leverage multimodal data, most efforts focus on the interaction between visual and textual features. However, they ignore the syntactic structures of the text during the interaction, where all textual components are intertwined, resulting in ambiguous vision-language alignment. In this paper, we improve the multimodal interaction by DECOUPLING the interweave. Specifically, we train a lightweight subject perceptron, which extracts the subject part from the input text. Then, the subject and text features are fed into two parallel branches to interact with visual features. This enables us to perform subject-aware and context-aware interactions, respectively, thus encouraging more explicit and discriminative feature embedding and alignment. Moreover, we find the decoupled architecture also facilitates incorporating the vision-language pre-trained alignment into RVOS, further improving the segmentation performance. Experimental results on all RVOS benchmark datasets demonstrate the superiority of our proposed method over the state-of-the-arts. The code of our method is available at:https://github.com/gaomingqi/dmformer.

Multi-class segmentation of cardiac magnetic resonance (CMR) images seeks a separation of data into anatomical components with known structure and configuration. The most popular CNN-based methods are optimised using pixel wise loss functions, ignorant of the spatially extended features that characterise anatomy. Therefore, whilst sharing a high spatial overlap with the ground truth, inferred CNN-based segmentations can lack coherence, including spurious connected components, holes and voids. Such results are implausible, violating anticipated anatomical topology. In response, (single-class) persistent homology-based loss functions have been proposed to capture global anatomical features. Our work extends these approaches to the task of multi-class segmentation. Building an enriched topological description of all class labels and class label pairs, our loss functions make predictable and statistically significant improvements in segmentation topology using a CNN-based post-processing framework. We also present (and make available) a highly efficient implementation based on cubical complexes and parallel execution, enabling practical application within high resolution 3D data for the first time. We demonstrate our approach on 2D short axis and 3D whole heart CMR segmentation, advancing a detailed and faithful analysis of performance on two publicly available datasets.

Combinatorial Optimisation problems arise in several application domains and are often formulated in terms of graphs. Many of these problems are NP-hard, but exact solutions are not always needed. Several heuristics have been developed to provide near-optimal solutions; however, they do not typically scale well with the size of the graph. We propose a low-complexity approach for identifying a (possibly much smaller) subgraph of the original graph where the heuristics can be run in reasonable time and with a high likelihood of finding a global near-optimal solution. The core component of our approach is LeNSE, a reinforcement learning algorithm that learns how to navigate the space of possible subgraphs using an Euclidean subgraph embedding as its map. To solve CO problems, LeNSE is provided with a discriminative embedding trained using any existing heuristics using only on a small portion of the original graph. When tested on three problems (vertex cover, max-cut and influence maximisation) using real graphs with up to $10$ million edges, LeNSE identifies small subgraphs yielding solutions comparable to those found by running the heuristics on the entire graph, but at a fraction of the total run time. Code for the experiments is available in the public GitHub repo at https://github.com/davidireland3/LeNSE.

Abstract Mastering robotic manipulation skills through reinforcement learning (RL) typically requires the design of shaped reward functions. Recent developments in this area have demonstrated that using sparse rewards, i.e. rewarding the agent only when the task has been successfully completed, can lead to better policies. However, state-action space exploration is more difficult in this case. Recent RL approaches to learning with sparse rewards have leveraged high-quality human demonstrations for the task, but these can be costly, time consuming or even impossible to obtain. In this paper, we propose a novel and effective approach that does not require human demonstrations. We observe that every robotic manipulation task could be seen as involving a locomotion task from the perspective of the object being manipulated, i.e. the object could learn how to reach a target state on its own. In order to exploit this idea, we introduce a framework whereby an object locomotion policy is initially obtained using a realistic physics simulator. This policy is then used to generate auxiliary rewards, called simulated locomotion demonstration rewards (SLDRs), which enable us to learn the robot manipulation policy. The proposed approach has been evaluated on 13 tasks of increasing complexity, and can achieve higher success rate and faster learning rates compared to alternative algorithms. SLDRs are especially beneficial for tasks like multi-object stacking and non-rigid object manipulation.

Training agents to autonomously control anthropomorphic robotic hands has the potential to lead to systems capable of performing a multitude of complex manipulation tasks in unstructured and uncertain environments. In this work, we first introduce a suite of challenging simulated manipulation tasks where current reinforcement learning and trajectory optimisation techniques perform poorly. These include environments where two simulated hands have to pass or throw objects between each other, as well as an environment where the agent must learn to spin a long pen between its fingers. We then introduce a simple trajectory optimisation algorithm that performs significantly better than existing methods on these environments. Finally, on the most challenging “PenSpin" task, we combine sub-optimal demonstrations generated through trajectory optimisation with off-policy reinforcement learning, obtaining performance that far exceeds either of these approaches individually. Videos of all of our results are available at: https://dexterous-manipulation.github.io

Learning with sparse rewards remains a significant challenge in reinforcement learning (RL), especially when the aim is to train a policy capable of achieving multiple different goals. To date, the most successful approaches for dealing with multi-goal, sparse reward environments have been model-free RL algorithms. In this work we propose PlanGAN, a model-based algorithm specifically designed for solving multi-goal tasks in environments with sparse rewards. Our method builds on the fact that any trajectory of experience collected by an agent contains useful information about how to achieve the goals observed during that trajectory. We use this to train an ensemble of conditional generative models (GANs) to generate plausible trajectories that lead the agent from its current state towards a specified goal. We then combine these imagined trajectories into a novel planning algorithm in order to achieve the desired goal as efficiently as possible. The performance of PlanGAN has been tested on a number of robotic navigation/manipulation tasks in comparison with a range of model-free reinforcement learning baselines, including Hindsight Experience Replay. Our studies indicate that PlanGAN can achieve comparable performance whilst being around 4-8 times more sample efficient.

Abstract Deep reinforcement learning algorithms have recently been used to train multiple interacting agents in a centralised manner whilst keeping their execution decentralised. When the agents can only acquire partial observations and are faced with tasks requiring coordination and synchronisation skills, inter-agent communication plays an essential role. In this work, we propose a framework for multi-agent training using deep deterministic policy gradients that enables concurrent, end-to-end learning of an explicit communication protocol through a memory device. During training, the agents learn to perform read and write operations enabling them to infer a shared representation of the world. We empirically demonstrate that concurrent learning of the communication device and individual policies can improve inter-agent coordination and performance in small-scale systems. Our experimental results show that the proposed method achieves superior performance in scenarios with up to six agents. We illustrate how different communication patterns can emerge on six different tasks of increasing complexity. Furthermore, we study the effects of corrupting the communication channel, provide a visualisation of the time-varying memory content as the underlying task is being solved and validate the building blocks of the proposed memory device through ablation studies.

Abstract The automatic analysis of ultrasound sequences can substantially improve the efficiency of clinical diagnosis. This article presents an attempt to automate the challenging task of measuring the vascular diameter of the fetal abdominal aorta from ultrasound images. We propose a neural network architecture consisting of three blocks: a convolutional neural network (CNN) for the extraction of imaging features, a convolution gated recurrent unit (C-GRU) for exploiting the temporal redundancy of the signal, and a regularized loss function, calledCyclicLoss, to impose our prior knowledge about the periodicity of the observed signal. The solution is investigated with a cohort of 25 ultrasound sequences acquired during the third-trimester pregnancy check, and with 1000 synthetic sequences. In the extraction of features, it is shown that a shallow CNN outperforms two other deep CNNs with both the real and synthetic cohorts, suggesting that echocardiographic features are optimally captured by a reduced number of CNN layers. The proposed architecture, working with the shallow CNN, reaches an accuracy substantially superior to previously reported methods, providing an average reduction of the mean squared error from 0.31 (state-of-the-art) to 0.09 $$\mathrm{mm}^2$$ mm2 , and a relative error reduction from 8.1 to 5.3%. The mean execution speed of the proposed approach of 289 frames per second makes it suitable for real-time clinical use.

Machine learning approaches hold great potential for the automated detection of lung nodules on chest radiographs, but training algorithms requires very large amounts of manually annotated radiographs, which are difficult to obtain. The increasing availability of PACS (Picture Archiving and Communication System), is laying the technological foundations needed to make available large volumes of clinical data and images from hospital archives. Binary labels indicating whether a radiograph contains a pulmonary lesion can be extracted at scale, using natural language processing algorithms. In this study, we propose two novel neural networks for the detection of chest radiographs containing pulmonary lesions. Both architectures make use of a large number of weakly-labelled images combined with a smaller number of manually annotated x-rays. The annotated lesions are used during training to deliver a type of visual attention feedback informing the networks about their lesion localisation performance. The first architecture extracts saliency maps from high-level convolutional layers and compares the inferred position of a lesion against the true position when this information is available; a localisation error is then back-propagated along with the softmax classification error. The second approach consists of a recurrent attention model that learns to observe a short sequence of smaller image portions through reinforcement learning; the reward function penalises the exploration of areas, within an image, that are unlikely to contain nodules. Using a repository of over 430,000 historical chest radiographs, we present and discuss the proposed methods over related architectures that use either weakly-labelled or annotated images only.

Multiplex networks, a special type of multilayer networks, are increasingly applied in many domains ranging from social media analytics to biology. A common task in these applications concerns the detection of community structures. Many existing algorithms for community detection in multiplexes attempt to detect communities which are shared by all layers. In this article we propose a community detection algorithm, LART (Locally Adaptive Random Transitions), for the detection of communities that are shared by either some or all the layers in the multiplex. The algorithm is based on a random walk on the multiplex, and the transition probabilities defining the random walk are allowed to depend on the local topological similarity between layers at any given node so as to facilitate the exploration of communities across layers. Based on this random walk, a node dissimilarity measure is derived and nodes are clustered based on this distance in a hierarchical fashion. We present experimental results using networks simulated under various scenarios to showcase the performance of LART in comparison to related community detection algorithms.

This paper presents a new data-driven classification pipeline for discriminating two groups of individuals based on the medical images of their brain.The algorithm combines deformation-based morphometry and penalised linear discriminant analysis with resampling.The method is based on sparse representation of the original brain images using deformation logarithms reflecting the differences in the brain in comparison to the normal template anatomy.The sparse data enables efficient data reduction and classification via the penalised linear discriminant analysis with resampling.The classification accuracy obtained in an experiment with magnetic resonance brain images of first episode schizophrenia patients and healthy controls is comparable to the related state-of-the-art studies. I. INTRODUCTIONhe last two decades have witnessed an explosive growth in the ability to "understand the human brain"a key to progress in neuroscience, to promote and protect brain health, and to develop treatments for restoring, regenerating, and repairing diseased brain functions.The motivations for that are clear: as populations inevitably grow older, mental disorders will increase dramatically -implying economic and social implications.The identification, characterization and validation of biomarkers for the major mental disorders would facilitate accurate prediction of disease risk, course, and therapeutic responses and ultimately lead to knowledge-based treatment and preventive strategies.Computational neuroanatomy is a growing field of powerful applications of imaging modalities and computational techniques in neuroscience.It promises an automated methodology to characterize neuroanatomical configuration of structural magnetic resonance imaging (MRI) brain scans.One of the crucial techniques in this methodology is image registration.Its task is to find a spatial transformation which maps each point of an image onto its

MOTIVATION: Within any given tissue, gene expression levels can vary extensively among individuals. Such heterogeneity can be caused by genetic and epigenetic variability and may contribute to disease. The abundance of experimental data now enables the identification of features of gene expression profiles that are shared across tissues and those that are tissue-specific. While most current research is concerned with characterizing differential expression by comparing mean expression profiles across tissues, it is believed that a significant difference in a gene expression's variance across tissues may also be associated with molecular mechanisms that are important for tissue development and function. RESULTS: We propose a sparse multi-view matrix factorization (sMVMF) algorithm to jointly analyse gene expression measurements in multiple tissues, where each tissue provides a different 'view' of the underlying organism. The proposed methodology can be interpreted as an extension of principal component analysis in that it provides the means to decompose the total sample variance in each tissue into the sum of two components: one capturing the variance that is shared across tissues and one isolating the tissue-specific variances. sMVMF has been used to jointly model mRNA expression profiles in three tissues obtained from a large and well-phenotyped twins cohort, TwinsUK. Using sMVMF, we are able to prioritize genes based on whether their variation patterns are specific to each tissue. Furthermore, using DNA methylation profiles available, we provide supporting evidence that adipose-specific gene expression patterns may be driven by epigenetic effects. AVAILABILITY AND IMPLEMENTATION: Python code is available at http://wwwf.imperial.ac.uk/~gmontana/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: In cancer research, the comparison of gene expression or DNA methylation networks inferred from healthy controls and patients can lead to the discovery of biological pathways associated to the disease. As a cancer progresses, its signalling and control networks are subject to some degree of localised re-wiring. Being able to detect disrupted interaction patterns induced by the presence or progression of the disease can lead to the discovery of novel molecular diagnostic and prognostic signatures. Currently there is a lack of scalable statistical procedures for two-network comparisons aimed at detecting localised topological differences. RESULTS: We propose the dGHD algorithm, a methodology for detecting differential interaction patterns in two-network comparisons. The algorithm relies on a statistic, the Generalised Hamming Distance (GHD), for assessing the degree of topological difference between networks and evaluating its statistical significance. dGHD builds on a non-parametric permutation testing framework but achieves computationally efficiency through an asymptotic normal approximation. CONCLUSIONS: We show that the GHD is able to detect more subtle topological differences compared to a standard Hamming distance between networks. This results in the dGHD algorithm achieving high performance in simulation studies as measured by sensitivity and specificity. An application to the problem of detecting differential DNA co-methylation subnetworks associated to ovarian cancer demonstrates the potential benefits of the proposed methodology for discovering network-derived biomarkers associated with a trait of interest.

MOTIVATION: High-throughput profiling in biological research has resulted in the availability of a wealth of data cataloguing the genetic, epigenetic and transcriptional states of cells. These data could yield discoveries that may lead to breakthroughs in the diagnosis and treatment of human disease, but require statistical methods designed to find the most relevant patterns from millions of potential interactions. Aberrant DNA methylation is often a feature of cancer, and has been proposed as a therapeutic target. However, the relationship between DNA methylation and gene expression remains poorly understood. RESULTS: We propose Network-sparse Reduced-Rank Regression (NsRRR), a multivariate regression framework capable of using prior biological knowledge expressed as gene interaction networks to guide the search for associations between gene expression and DNA methylation signatures. We use simulations to show the advantage of our proposed model in terms of variable selection accuracy over alternative models that do not use prior network information. We discuss an application of NsRRR to The Cancer Genome Atlas datasets on primary ovarian tumours. AVAILABILITY AND IMPLEMENTATION: R code implementing the NsRRR model is available at http://www2.imperial.ac.uk/∼gmontana CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Due to rapid technological advances, a wide range of different measurements can be obtained from a given biological sample including single nucleotide polymorphisms, copy number variation, gene expression levels, DNA methylation and proteomic profiles. Each of these distinct measurements provides the means to characterize a certain aspect of biological diversity, and a fundamental problem of broad interest concerns the discovery of shared patterns of variation across different data types. Such data types are heterogeneous in the sense that they represent measurements taken at different scales or represented by different data structures. RESULTS: We propose a distance-based statistical test, the generalized RV (GRV) test, to assess whether there is a common and non-random pattern of variability between paired biological measurements obtained from the same random sample. The measurements enter the test through the use of two distance measures, which can be chosen to capture a particular aspect of the data. An approximate null distribution is proposed to compute P-values in closed-form and without the need to perform costly Monte Carlo permutation procedures. Compared with the classical Mantel test for association between distance matrices, the GRV test has been found to be more powerful in a number of simulation settings. We also demonstrate how the GRV test can be used to detect biological pathways in which genetic variability is associated to variation in gene expression levels in an ovarian cancer sample, and present results obtained from two independent cohorts. AVAILABILITY: R code to compute the GRV test is freely available from http://www2.imperial.ac.uk/∼gmontana

MOTIVATION: Multivariate quantitative traits arise naturally in recent neuroimaging genetics studies, in which both structural and functional variability of the human brain is measured non-invasively through techniques such as magnetic resonance imaging (MRI). There is growing interest in detecting genetic variants associated with such multivariate traits, especially in genome-wide studies. Random forests (RFs) classifiers, which are ensembles of decision trees, are amongst the best performing machine learning algorithms and have been successfully employed for the prioritisation of genetic variants in case-control studies. RFs can also be applied to produce gene rankings in association studies with multivariate quantitative traits, and to estimate genetic similarities measures that are predictive of the trait. However, in studies involving hundreds of thousands of SNPs and high-dimensional traits, a very large ensemble of trees must be inferred from the data in order to obtain reliable rankings, which makes the application of these algorithms computationally prohibitive. RESULTS: We have developed a parallel version of the RF algorithm for regression and genetic similarity learning tasks in large-scale population genetic association studies involving multivariate traits, called PaRFR (Parallel Random Forest Regression). Our implementation takes advantage of the MapReduce programming model and is deployed on Hadoop, an open-source software framework that supports data-intensive distributed applications. Notable speed-ups are obtained by introducing a distance-based criterion for node splitting in the tree estimation process. PaRFR has been applied to a genome-wide association study on Alzheimer's disease (AD) in which the quantitative trait consists of a high-dimensional neuroimaging phenotype describing longitudinal changes in the human brain structure. PaRFR provides a ranking of SNPs associated to this trait, and produces pair-wise measures of genetic proximity that can be directly compared to pair-wise measures of phenotypic proximity. Several known AD-related variants have been identified, including APOE4 and TOMM40. We also present experimental evidence supporting the hypothesis of a linear relationship between the number of top-ranked mutated states, or frequent mutation patterns, and an indicator of disease severity. AVAILABILITY: The Java codes are freely available at http://www2.imperial.ac.uk/~gmontana.

MOTIVATION: Metabolomics is the study of the complement of small molecule metabolites in cells, biofluids and tissues. Many metabolomic experiments are designed to compare changes observed over time under two experimental conditions or groups (e.g. a control and drug-treated group) with the goal of identifying discriminatory metabolites or biomarkers that characterize each condition. A common study design consists of repeated measurements taken on each experimental unit thus producing time courses of all metabolites. We describe a statistical framework for estimating time-varying metabolic profiles and their within-group variability and for detecting between-group differences. Specifically, we propose (i) a smoothing splines mixed effects (SME) model that treats each longitudinal measurement as a smooth function of time and (ii) an associated functional test statistic. Statistical significance is assessed by a non-parametric bootstrap procedure. RESULTS: The methodology has been extensively evaluated using simulated data and has been applied to real nuclear magnetic resonance spectroscopy data collected in a preclinical toxicology study as part of a larger project lead by the COMET (Consortium for Metabonomic Toxicology). Our findings are compatible with the previously published studies. AVAILABILITY: An R script is freely available for download at http://www2.imperial.ac.uk/~gmontana/sme.htm.

MOTIVATION: Time course gene expression experiments are performed to study time-varying changes in mRNA levels of thousands of genes. Statistical methods from functional data analysis (FDA) have recently gained popularity for modelling and exploring such time courses. Each temporal profile is treated as the realization of a smooth function of time, or curve, and the inferred curve becomes the basic unit of statistical analysis. The task of identifying genes with differential temporal profiles then consists of detecting statistically significant differences between curves, where such differences are commonly quantified by computing the area between the curves or the l₂ distance. RESULTS: We propose a general test statistic for detecting differences between gene curves, which only depends on a suitably chosen distance measure between them. The test makes use of a distance-based variance decomposition and generalizes traditional MANOVA tests commonly used for vectorial observations. We also introduce the visual l₂ distance, which is shown to capture shape-related differences in gene curves and is robust against time shifts, which would otherwise inflate the traditional l₂ distance. Other shape-related distances, such as the curvature, may carry biological significance. We have assessed the comparative performance of the test on realistically simulated datasets and applied it to human immune cell responses to bacterial infection over time.

Admixture mapping is a statistical methodology that detects genetic variants in recently admixed populations that are responsible for ethnic differences in disease risk. Three software packages are now available for admixture mapping and we provide a brief overview of the statistical methods and other principal features they implement.

In recent years, a very large variety of statistical methodologies, at various levels of complexity, have been put forward to analyse genotype data and detect genetic variations that may be responsible for increasing the susceptibility to disease. This review provides a concise account of a number of selected statistical methods for population-based association mapping, from single-marker tests of association to multi-marker data mining techniques for gene-gene interaction detection.

UNLABELLED: We have developed a simulation tool HapSim for the generation of haplotype data. The simulated haplotypes are such that their allele frequencies and linkage disequilibrium coefficients match exactly those estimated in a real sample. AVAILABILITY: The program is available as an R package and can be downloaded from http://cran.r-project.org/.