With the growing ubiquity of video content, efficient video analytics has become essential for applications such as surveillance, autonomous driving, and augmented reality. Yet, deploying video analytics models on resource-constrained edge devices and in low-bandwidth environments remains challenging. A dominant method for handling demanding video analytics tasks on edge devices has been to offload computation strategically from the edge device to servers. However, all prior solutions fail to offload under severely constrained, real-world network conditions (such as, a few-Mbps satellite network) due to the much higher data rates associated with video tasks. We introduce ApproxBit, a system to optimize shared edge-to-cloud processing for video analytics tasks; the two that we experiment with are video action recognition and video question answering. ApproxBit integrates an encoder within the video model, uses learned binary codes to effectively compress and offload data, and adaptively decides on the offloading point depending on the network bandwidth. ApproxBit’s adaptive and efficient data compression, which reduces the original feature map size by up to 2142.4 ×, makes it an ideal solution for video analytics on edge devices, especially with constrained networks. We evaluate ApproxBit on the two video tasks, across different model architectures (e.g., convolution- and Transformer-based) and multiple datasets (e.g., Something-Something-v2, Kinetics, and MSVD). Our results of latency and accuracy are superior over baselines: edge-only processing, server-only processing, DNN Surgery [ToCC ’23], full offloading of H.264-encoded videos, DeepCOD [SenSys ’20], neural video compression DCVC-FM [CVPR ’24], and LimitNet [MobiSys ’24]. We also demonstrate ApproxBit’s adaptivity to changing network conditions, and generalization in a real-world user study.

Semi-supervised semantic segmentation with consistency regularization capitalizes on unlabeled images to enhance the accuracy of pixel-level segmentation. Current consistency learning methods primarily rely on the consistency loss between pseudo-labels and unlabeled images, neglecting the information within the feature representations of the backbone encoder. Preserving maximum information in feature embeddings requires achieving the alignment and uniformity objectives, as widely studied. To address this, we present SWSeg, a semi-supervised semantic segmentation algorithm that optimizes alignment and uniformity using the Sliced-Wasserstein Distance (SWD), and rigorously and empirically proves this connection. We further resolve the computational issues associated with conventional Monte Carlo-based SWD by implementing a Gaussian-approximated variant, which not only maintains the alignment and uniformity objectives but also improves training efficiency. We evaluate SWSeg on the PASCAL VOC 2012, Cityscapes, and ADE20K datasets, outshining supervised baselines in mIoU by up to 11.8%, 8.9%, and 8.2%, respectively, given an equivalent number of labeled samples. Further, SWSeg surpasses state-of-the-art methods in multiple settings across these three datasets. Our extensive ablation studies confirm the optimization of the uniformity and alignment objectives of the feature representations.

We present SKALD, a multi-shot video assembly method that constructs coherent video sequences from candidate shots with minimal reliance on text. Central to our approach is the Learned Clip Assembly (LCA) score, a learning-based metric that measures temporal and semantic relationships between shots to quantify narrative coherence. We tackle the exponential complexity of combining multiple shots with an efficient beam-search algorithm guided by the LCA score. To train our model effectively with limited human annotations, we propose two tasks for the LCA encoder: Shot Coherence Learning, which uses contrastive learning to distinguish coherent and incoherent sequences, and Feature Regression, which converts these learned representations into a real-valued coherence score. We develop two variants: a base SKALD model that relies solely on visual coherence and SKALD-text, which integrates auxiliary text information when available. Experiments on the VSPD and our curated MSV3C datasets show that SKALD achieves an improvement of up to 48.6% in IoU and a 43% speedup over the state-of-the-art methods. A user study further validates our approach, with 45% of participants favoring SKALD-assembled videos, compared to 22% preferring text-based assembly methods.

Multimodal Large Language Models (MLLMs) have shown impressive capabilities in visual reasoning, yet come with substantial computational cost, limiting their deployment in resource-constrained settings. Despite recent effort on improving the efficiency of MLLMs, prior solutions fall short in responding to varying runtime conditions, in particular changing resource availability (e.g., contention due to the execution of other programs on the device). To bridge this gap, we introduce AdaLLaVA, an adaptive inference framework that learns to dynamically reconfigure operations in an MLLM during inference, accounting for the input data and a latency budget. We conduct extensive experiments across benchmarks involving question-answering, reasoning, and hallucination. Our results show that AdaLLaVA effectively adheres to input latency budget, achieving varying accuracy and latency tradeoffs at runtime. Further, we demonstrate that AdaLLaVA adapts to both input latency and content, can be integrated with token selection for enhanced efficiency, and generalizes across MLLMs. Our project webpage with code release is at https://zhuoyan-xu.github.io/ada-llava/.

Efficient 3D perception is critical for autonomous systems—self-driving vehicles, drones—to navigate safely in dynamic environments. Accurate 3D object detection from LiDAR data must handle irregular, high-volume point clouds, variable latency from contention and scene complexity, and tight embedded GPU constraints. Balancing accuracy and latency under dynamic conditions is crucial, yet existing frameworks like Chanakya [NeurIPS '23], LiteReconfig [EuroSys '22], and AdaScale [MLSys '19] struggle with the unique demands of 3D detection. We present Agile3D, the first adaptive 3D system integrating a cross-model Multi-branch Execution Framework (MEF) and a Contention- and Content-Aware Reinforcement Learning-based controller (CARL). CARL dynamically selects the optimal execution branch using five novel MEF control knobs: encoding format, spatial resolution, spatial encoding, 3D feature extractor, and detection head. CARL uses supervised training for stable initial policies, then Direct Preference Optimization (DPO) to finetune branch selection without hand-crafted rewards, presenting the first application of DPO to branch scheduling in 3D detection. Comprehensive evaluations show that Agile3D achieves state-of-the-art performance, maintaining high accuracy across varying hardware contention levels and 100-500 ms latency budgets. On NVIDIA Orin and Xavier GPUs, it consistently leads the Pareto frontier, outperforming existing methods for efficient 3D detection.

The target localization primitive is used for detecting and locating an adverse event called a target in a geographic area. This versatile primitive is applicable in the physical security domain (e.g., detecting intruders in an area) or for disaster preemption, such as detecting ignition events of forest fires. Prior systems implemented this primitive over large areas by deploying a network of sensor devices, which detect changes in a specific physical parameter like pressure or temperature induced by a target. However, these systems are not designed for use in adverse environments where one or more sensors can behave in a faulty manner. While many algorithms in the distributed systems literature can be naively used to implement target localization in a fault-tolerant manner, these approaches are energy-intensive as they use computationally expensive cryptographic operations not appropriate for resource-constrained sensors. We present SENSORBFT, an energy-efficient, fault-tolerant approach for target localization. SENSORBFT uses a novel asynchronous approximate agreement protocol that enables correct sensors to achieve an approximate consensus in the presence of faulty sensors. Sensors fulfill their energy budgets by tuning the precision and accuracy of localization, where precision is the difference between honest sensors' outputs and accuracy is the difference between an honest sensor's output and the target's true location. In optimal scenarios, this protocol reduces communication from$O$($n$3) to$O$($n$2) messages per round, where$n$is the number of sensors sharing coverage over a piece of area. In a sensor testbed with$n$= 19 sensors, SENSORBFT consumes 2/5 th the energy consumed by existing solutions for a minor 2% loss in accuracy, significantly enhancing efficiency and coverage.

Federated learning—multi-party, distributed learning in a decentralized environment—is vulnerable to model poisoning attacks, more so than centralized learning. This is because malicious clients can collude and send in carefully tailored model updates to make the global model inaccurate. This motivated the development of Byzantine-resilient federated learning algorithms, such as Krum, Bulyan, FABA, and FoolsGold. However, a recently developed untargeted model poisoning attack showed that all prior defenses can be bypassed. The attack uses the intuition that simply by changing the sign of the gradient updates that the optimizer is computing, for a set of malicious clients, a model can be diverted from the optima to increase the test error rate. In this work, we develop FLAIR—a defense against this directed deviation attack (DDA), a state-of-the-art model poisoning attack. FLAIR is based on our intuition that in federated learning, certain patterns of gradient flips are indicative of an attack. This intuition is remarkably stable across different learning algorithms, models, and datasets. FLAIR assigns reputation scores to the participating clients based on their behavior during the training phase and then takes a weighted contribution of the clients. We show that where the existing defense baselines of FABA [IJCAI ’19], FoolsGold [Usenix ’20], and FLTrust [NDSS ’21] fail when 20-30% of the clients are malicious, FLAIR provides byzantine-robustness upto a malicious client percentage of 45%. We also show that FLAIR provides robustness against even a white-box version of DDA.

Efficient and adaptive computer vision systems have been proposed to make computer vision tasks, such as image classification and object detection, optimized for embedded or mobile devices. These solutions, quite recent in their origin, focus on optimizing the model (a deep neural network) or the system by designing an adaptive system with approximation knobs. Despite several recent efforts, we show that existing solutions suffer from two major drawbacks. First , while mobile devices or systems-on-chips usually come with limited resources including battery power, most systems do not consider the energy consumption of the models during inference. Second , they do not consider the interplay between the three metrics of interest in their configurations, namely, latency, accuracy, and energy. In this work, we propose an efficient and adaptive video object detection system— Virtuoso , which is jointly optimized for accuracy, energy efficiency, and latency. Underlying Virtuoso is a multi-branch execution kernel that is capable of running at different operating points in the accuracy-energy-latency axes, and a lightweight runtime scheduler to select the best fit execution branch to satisfy the user requirement. We position this work as a first step in understanding the suitability of various object detection kernels on embedded boards in the accuracy-latency-energy axes, opening the door for further development in solutions customized to embedded systems and for benchmarking such solutions. Virtuoso is able to achieve up to 286 FPS on the NVIDIA Jetson AGX Xavier board, which is up to 45× faster than the baseline EfficientDet D3 and 15× faster than the baseline EfficientDet D0. In addition, we also observe up to 97.2% energy reduction using Virtuoso compared to the baseline YOLO (v3)—a widely used object detector designed for mobiles. To fairly compare with Virtuoso , we benchmark 15 state-of-the-art or widely used protocols, including Faster R-CNN (FRCNN) [NeurIPS’15], YOLO v3 [CVPR’16], SSD [ECCV’16], EfficientDet [CVPR’20], SELSA [ICCV’19], MEGA [CVPR’20], REPP [IROS’20], FastAdapt [EMDL’21], and our in-house adaptive variants of FRCNN+, YOLO+, SSD+, and EfficientDet+ (our variants have enhanced efficiency for mobiles). With this comprehensive benchmark, Virtuoso has shown superiority to all the above protocols, leading the accuracy frontier at every efficiency level on NVIDIA Jetson mobile GPUs. Specifically, Virtuoso has achieved an accuracy of 63.9%, which is more than 10% higher than some of the popular object detection models, FRCNN at 51.1% and YOLO at 49.5%.

Several recent works seek to create lightweight deep net-works for video object detection on mobiles. We observe that many existing detectors, previously deemed computationally costly for mobiles, intrinsically support adaptive inference, and offer a multi-branch object detection frame-work (MBODF). Here, an MBODF is referred to as a so-lution that has many execution branches and one can dy-namically choose from among them at inference time to sat-isfy varying latency requirements (e.g. by varying resolution of an input frame). In this paper, we ask, and answer, the wide-ranging question across all MBODFs: How to expose the right set of execution branches and then how to sched-ule the optimal one at inference time? In addition, we un-cover the importance of making a content-aware decision on which branch to run, as the optimal one is conditioned on the video content. Finally, we explore a content-aware scheduler, an Oracle one, and then a practical one, leveraging various lightweight feature extractors. Our evaluation shows that layered on Faster R-CNN-based MBODF, compared to 7 baselines, our Smartadapt achieves a higher Pareto optimal curve in the accuracy-vs-latency space for the ILSVRC VID dataset.

An adaptive video object detection system selects different execution paths at runtime, based on video content and available resources, so as to maximize accuracy under a target latency objective (e.g., 30 frames per second). Such a system is well suited to mobile devices with limited computing resources, and often running multiple contending applications. Existing solutions suffer from two major drawbacks. First, collecting feature values to decide on an execution branch is expensive. Second, there is a switching overhead for transitioning between branches and this overhead depends on the transition pair. LiteReconfig, an efficient and adaptive video object detection framework, addresses these challenges. LiteReconfig features a cost-benefit analyzer to decide which features to use, and which execution branch to run, at inference time. Furthermore, LiteReconfig has a content-aware accuracy prediction model, to select an execution branch tailored for frames in a video stream. We demonstrate that LiteReconfig achieves significantly improved accuracy under a set of varying latency objectives than existing systems, while maintaining up to 50 fps on an NVIDIA AGX Xavier board. Our code, with DOI, is available at https://doi.org/10.5281/zenodo.6345733.

A common workflow for single-cell RNA-sequencing (sc-RNA-seq) data analysis is to orchestrate a three-step pipeline. First, conduct a dimension reduction of the input cell profile matrix; second, cluster the cells in the latent space; and third, extract the "gene panels" that distinguish a certain cluster from others. This workflow has the primary drawback that the three steps are performed independently, neglecting the dependencies among the steps and among the marker genes or gene panels. In our system, KRATOS, we alter the three-step workflow to a two-step one, where we jointly optimize the first two steps and add the third (interpretability) step to form an integrated sc-RNA-seq analysis pipeline. We show that the more compact workflow of KRATOS extracts marker genes that can better discriminate the target cluster, distilling underlying mechanisms guiding cluster membership. In doing so, KRATOS is significantly better than the two SOTA baselines we compare against, specifically 5.62% superior to Global Counterfactual Explanation (GCE) [ICML-20], and 3.31% better than Adversarial Clustering Explanation (ACE) [ICML-21], measured by the AUROC of a kernel-SVM classifier. We opensource our code and datasets here: https://github.com/icanforce/single-cell-genomics-kratos.

BACKGROUND: Sequencing technologies are prone to errors, making error correction (EC) necessary for downstream applications. EC tools need to be manually configured for optimal performance. We find that the optimal parameters (e.g., k-mer size) are both tool- and dataset-dependent. Moreover, evaluating the performance (i.e., Alignment-rate or Gain) of a given tool usually relies on a reference genome, but quality reference genomes are not always available. We introduce Lerna for the automated configuration of k-mer-based EC tools. Lerna first creates a language model (LM) of the uncorrected genomic reads, and then, based on this LM, calculates a metric called the perplexity metric to evaluate the corrected reads for different parameter choices. Next, it finds the one that produces the highest alignment rate without using a reference genome. The fundamental intuition of our approach is that the perplexity metric is inversely correlated with the quality of the assembly after error correction. Therefore, Lerna leverages the perplexity metric for automated tuning of k-mer sizes without needing a reference genome. RESULTS: First, we show that the best k-mer value can vary for different datasets, even for the same EC tool. This motivates our design that automates k-mer size selection without using a reference genome. Second, we show the gains of our LM using its component attention-based transformers. We show the model's estimation of the perplexity metric before and after error correction. The lower the perplexity after correction, the better the k-mer size. We also show that the alignment rate and assembly quality computed for the corrected reads are strongly negatively correlated with the perplexity, enabling the automated selection of k-mer values for better error correction, and hence, improved assembly quality. We validate our approach on both short and long reads. Additionally, we show that our attention-based models have significant runtime improvement for the entire pipeline-18[Formula: see text] faster than previous works, due to parallelizing the attention mechanism and the use of JIT compilation for GPU inferencing. CONCLUSION: Lerna improves de novo genome assembly by optimizing EC tools. Our code is made available in a public repository at: https://github.com/icanforce/lerna-genomics .

Videos take a lot of time to transport over the network, hence running analytics on the live video on embedded or mobile devices has become an important system driver. Considering such devices, e.g., surveillance cameras or AR/VR gadgets, are resource constrained, although there has been significant work in creating lightweight deep neural networks (DNNs) for such clients, none of these can adapt to changing runtime conditions, e.g., changes in resource availability on the device, the content characteristics, or requirements from the user. In this article, we introduce ApproxNet, a video object classification system for embedded or mobile clients. It enables novel dynamic approximation techniques to achieve desired inference latency and accuracy trade-off under changing runtime conditions. It achieves this by enabling two approximation knobs within a single DNN model rather than creating and maintaining an ensemble of models, e.g., MCDNN [MobiSys-16]. We show that ApproxNet can adapt seamlessly at runtime to these changes, provides low and stable latency for the image and video frame classification problems, and shows the improvement in accuracy and latency over ResNet [CVPR-16], MCDNN [MobiSys-16], MobileNets [Google-17], NestDNN [MobiCom-18], and MSDNet [ICLR-18].

BACKGROUND: MicroRNAs (miRNAs) function in post-transcriptional regulation of gene expression by binding to target messenger RNAs (mRNAs). Because of the key part that miRNAs play, understanding the correct regulatory role of miRNAs in diverse patho-physiological conditions is of great interest. Although it is known that miRNAs act combinatorially to regulate genes, precise identification of miRNA-gene interactions and their specific functional roles in regulatory comodules remains a challenge. We developed THEIA, an effective method for simultaneously predicting miRNA-gene interactions and regulatory comodules, which group functionally related miRNAs and genes via non-negative matrix factorization (NMF). RESULTS: We apply THEIA to RNA sequencing data from breast invasive carcinoma samples and demonstrate its effectiveness in discovering biologically significant regulatory comodules that are significantly enriched in spatial miRNA clusters, biological pathways, and various cancers. CONCLUSIONS: THEIA is a theoretically rigorous optimization algorithm that simultaneously predicts the strength and direction (i.e., up-regulation or down-regulation) of the effect of modules of miRNAs on a gene. We posit that if THEIA is capable of recovering known clusters of genes and miRNA, then the clusters found by our method not previously identified by literature are also likely to have biological significance. We believe that these novel regulatory comodules found by our method will be a springboard for further research into the specific functional roles of these new functional ensembles of miRNAs and genes,especially those related to diseases like breast cancer.

The recent advancement of the Internet of Things (IoT) enables the possibility of data collection from diverse environments using IoT devices. However, despite the rapid advancement of low-power communication technologies, the deployment of IoT networks still faces many challenges. In this article, we propose a hybrid, low-power, wide-area network (LPWAN) structure that can achieve wide-area communication coverage and low-power consumption on IoT devices by utilizing both sub-GHz long-range radio and 2.4-GHz short-range radio. Specifically, we constructed a low-power mesh network with LoRa, a physical-layer standard that can provide long-range (kilometers) point-to-point communication using custom time-division multiple access (TDMA). Furthermore, we extended the capabilities of the mesh network by enabling ANT, an ultralow-power, short-range communication protocol to satisfy data collection in dense device deployments. Third, we demonstrate the performance of the hybrid network with two real-world deployments at the Purdue University campus and at the university-owned farm. The results suggest that both networks have superior advantages in terms of cost, coverage, and power consumption vis-à-vis other IoT solutions, like LoRaWAN.

With diverse IoT workloads, placing compute and analytics close to where data is collected is becoming increasingly important. We seek to understand what is the performance and the cost implication of running analytics on IoT data at the various available platforms. These workloads can be compute-light, such as outlier detection on sensor data, or compute-intensive, such as object detection from video feeds obtained from drones. In our paper, JANUS, we profile the performance/$ and the compute versus communication cost for a compute-light IoT workload and a compute-intensive IoT workload. In addition, we also look at the pros and cons of some of the proprietary deep-learning object detection packages, such as Amazon Rekognition, Google Vision, and Azure Cognitive Services, to contrast with open-source and tunable solutions, such as Faster R-CNN (FRCNN). We find that AWS IoT Greengrass delivers at least 2X lower latency and 1.25X lower cost compared to all other cloud platforms for the compute-light outlier detection workload. For the compute-intensive streaming video analytics task, an opensource solution to object detection running on cloud VMs saves on dollar costs compared to proprietary solutions provided by Amazon, Microsoft, and Google, but loses out on latency (up to 6X). If it runs on a low-powered edge device, the latency is up to 49X lower.

Advanced video analytic systems, including scene classification and object detection, have seen widespread success in various domains such as smart cities and autonomous systems. With an evolution of heterogeneous client devices, there is incentive to move these heavy video analytics workloads from the cloud to mobile devices for low latency and real-time processing and to preserve user privacy. However, most video analytic systems are heavyweight and are trained offline with some pre-defined latency or accuracy requirements. This makes them unable to adapt at runtime in the face of three types of dynamism --- the input video characteristics change, the amount of compute resources available on the node changes due to co-located applications, and the user's latency-accuracy requirements change. In this paper we introduce ApproxDet, an adaptive video object detection framework for mobile devices to meet accuracy-latency requirements in the face of changing content and resource contention scenarios. To achieve this, we introduce a multi-branch object detection kernel, which incorporates a data-driven modeling approach on the performance metrics, and a latency SLA-driven scheduler to pick the best execution branch at runtime. We evaluate ApproxDet on a large benchmark video dataset and compare quantitatively to AdaScale and YOLOv3. We find that ApproxDet is able to adapt to a wide variety of contention and content characteristics and outshines all baselines, e.g., it achieves 52% lower latency and 11.1% higher accuracy over YOLOv3. Our software is open-sourced at https://github.com/purdue-dcsl/ApproxDet.

Federation is a popular concept in building distributed cyberinfrastructures, whereby computational resources are provided by multiple organizations through a unified portal, decreasing the complexity of moving data back and forth among multiple organizations. Federation has been used in bioinformatics only to a limited extent, namely, federation of datastores, e.g. SBGrid Consortium for structural biology and Gene Expression Omnibus (GEO) for functional genomics. Here, we posit that it is important to federate both computational resources (CPU, GPU, FPGA, etc.) and datastores to support popular bioinformatics portals, with fast-increasing data volumes and increasing processing requirements. A prime example, and one that we discuss here, is in genomics and metagenomics. It is critical that the processing of the data be done without having to transport the data across large network distances. We exemplify our design and development through our experience with metagenomics-RAST (MG-RAST), the most popular metagenomics analysis pipeline. Currently, it is hosted completely at Argonne National Laboratory. However, through a recently started collaborative National Institutes of Health project, we are taking steps toward federating this infrastructure. Being a widely used resource, we have to move toward federation without disrupting 50 K annual users. In this article, we describe the computational tools that will be useful for federating a bioinformatics infrastructure and the open research challenges that we see in federating such infrastructures. It is hoped that our manuscript can serve to spur greater federation of bioinformatics infrastructures by showing the steps involved, and thus, allow them to scale to support larger user bases.

As technologies change, MG-RAST is adapting. Newly available software is being included to improve accuracy and performance. As a computational service constantly running large volume scientific workflows, MG-RAST is the right location to perform benchmarking and implement algorithmic or platform improvements, in many cases involving trade-offs between specificity, sensitivity and run-time cost. The work in [Glass EM, Dribinsky Y, Yilmaz P, et al. ISME J 2014;8:1-3] is an example; we use existing well-studied data sets as gold standards representing different environments and different technologies to evaluate any changes to the pipeline. Currently, we use well-understood data sets in MG-RAST as platform for benchmarking. The use of artificial data sets for pipeline performance optimization has not added value, as these data sets are not presenting the same challenges as real-world data sets. In addition, the MG-RAST team welcomes suggestions for improvements of the workflow. We are currently working on versions 4.02 and 4.1, both of which contain significant input from the community and our partners that will enable double barcoding, stronger inferences supported by longer-read technologies, and will increase throughput while maintaining sensitivity by using Diamond and SortMeRNA. On the technical platform side, the MG-RAST team intends to support the Common Workflow Language as a standard to specify bioinformatics workflows, both to facilitate development and efficient high-performance implementation of the community's data analysis tasks.

BACKGROUND: The data deluge can leverage sophisticated ML techniques for functionally annotating the regulatory non-coding genome. The challenge lies in selecting the appropriate classifier for the specific functional annotation problem, within the bounds of the hardware constraints and the model's complexity. In our system AIKYATAN, we annotate distal epigenomic regulatory sites, e.g., enhancers. Specifically, we develop a binary classifier that classifies genome sequences as distal regulatory regions or not, given their histone modifications' combinatorial signatures. This problem is challenging because the regulatory regions are distal to the genes, with diverse signatures across classes (e.g., enhancers and insulators) and even within each class (e.g., different enhancer sub-classes). RESULTS: We develop a suite of ML models, under the banner AIKYATAN, including SVM models, random forest variants, and deep learning architectures, for distal regulatory element (DRE) detection. We demonstrate, with strong empirical evidence, deep learning approaches have a computational advantage. Plus, convolutional neural networks (CNN) provide the best-in-class accuracy, superior to the vanilla variant. With the human embryonic cell line H1, CNN achieves an accuracy of 97.9% and an order of magnitude lower runtime than the kernel SVM. Running on a GPU, the training time is sped up 21x and 30x (over CPU) for DNN and CNN, respectively. Finally, our CNN model enjoys superior prediction performance vis-'a-vis the competition. Specifically, AIKYATAN-CNN achieved 40% higher validation rate versus CSIANN and the same accuracy as RFECS. CONCLUSIONS: Our exhaustive experiments using an array of ML tools validate the need for a model that is not only expressive but can scale with increasing data volumes and diversity. In addition, a subset of these datasets have image-like properties and benefit from spatial pooling of features. Our AIKYATAN suite leverages diverse epigenomic datasets that can then be modeled using CNNs with optimized activation and pooling functions. The goal is to capture the salient features of the integrated epigenomic datasets for deciphering the distal (non-coding) regulatory elements, which have been found to be associated with functional variants. Our source code will be made publicly available at: https://bitbucket.org/cellsandmachines/aikyatan.

BACKGROUND: MicroRNAs (miRNAs) are approximately 22-nucleotide long regulatory RNA that mediate RNA interference by binding to cognate mRNA target regions. Here, we present a distributed kernel SVM-based binary classification scheme to predict miRNA targets. It captures the spatial profile of miRNA-mRNA interactions via smooth B-spline curves. This is accomplished separately for various input features, such as thermodynamic and sequence-based features. Further, we use a principled approach to uniformly model both canonical and non-canonical seed matches, using a novel seed enrichment metric. Finally, we verify our miRNA-mRNA pairings using an Elastic Net-based regression model on TCGA expression data for four cancer types to estimate the miRNAs that together regulate any given mRNA. RESULTS: We present a suite of algorithms for miRNA target prediction, under the banner Avishkar, with superior prediction performance over the competition. Specifically, our final kernel SVM model, with an Apache Spark backend, achieves an average true positive rate (TPR) of more than 75 percent, when keeping the false positive rate of 20 percent, for non-canonical human miRNA target sites. This is an improvement of over 150 percent in the TPR for non-canonical sites, over the best-in-class algorithm. We are able to achieve such superior performance by representing the thermodynamic and sequence profiles of miRNA-mRNA interaction as curves, devising a novel seed enrichment metric, and learning an ensemble of miRNA family-specific kernel SVM classifiers. We provide an easy-to-use system for large-scale interactive analysis and prediction of miRNA targets. All operations in our system, namely candidate set generation, feature generation and transformation, training, prediction, and computing performance metrics are fully distributed and are scalable. CONCLUSIONS: We have developed an efficient SVM-based model for miRNA target prediction using recent CLIP-seq data, demonstrating superior performance, evaluated using ROC curves for different species (human or mouse), or different target types (canonical or non-canonical). We analyzed the agreement between the target pairings using CLIP-seq data and using expression data from four cancer types. To the best of our knowledge, we provide the first distributed framework for miRNA target prediction based on Apache Hadoop and Spark. AVAILABILITY: All source code and sample data are publicly available at https://bitbucket.org/cellsandmachines/avishkar. Our scalable implementation of kernel SVM using Apache Spark, which can be used to solve large-scale non-linear binary classification problems, is available at https://bitbucket.org/cellsandmachines/kernelsvmspark.

Motivation: Transcription-regulatory elements (TREs) are critical modulators of gene transcription, with promoters and enhancers being major hubs for gene regulation. Of these, enhancers are distal regulatory elements, typically short 50-150 basepairs DNA regions and are identified via the large numbers of genomic regions displaying enhancer-like chromatin features, such as histone modifications at regulatory sites. Given that enhancers operate at sites vastly removed from their cognate genes, their computational prediction is challenging. While ChIP-seq datasets have been used to map out parts of the epigenome, only a fraction of predicted enhancers have been found to be functionally active. Thus, recent methods, such as GRO-seq, have been used to elucidate functionally active enhancers. However, GRO-seq data is sparse, being experimentally complex, and often at insufficient resolution. Hence, in this paper, we evaluate the ability of ChIP-seq data to predict the eRNA activity, as given by GRO-seq, using as inputs the signatures of key histone modifications that are known to be predictive of enhancers. We implement two models with accurate performance for the entire range of eRNA activity, spanning 5 orders of magnitude, for the human embryonic stem cell type H1. We then perform a detailed comparison with a prior computational approach for (binary) enhancer activity classification. Results: Our best performing model is the Convolutional Neural Network (CNN)-based regression model, indicating that the epigenetic patterns are location invariant. We achieve an overall RMSE of 0.82 compared to the baseline's (ZeroR) 1.30. In the comparison with a prior approach, our model achieves AUC of 0.75 against the prior approach's 0.56.

High performance computing (HPC) applications, such as metagenomics and other big data systems, need to store and analyze huge volumes of semi-structured data. Such applications often rely on NoSQL-based datastores, and optimizing these databases is a challenging endeavor, with over 50 configuration parameters in Cassandra alone. As the application executes, database workloads can change rapidly from read-heavy to write-heavy ones, and a system tuned with a read-optimized configuration becomes suboptimal when the workload becomes write-heavy.

Breakthroughs in gene sequencing technologies have led to an exponential increase in the amount of genomic data. Efficient tools to rapidly process such large quantities of data are critical in the study of gene functions, diseases, evolution, and population variation. These tools are designed in an ad-hoc manner, and require extensive programmer effort to develop and optimize them. Often, such tools are written with the currently available data sizes in mind, and soon start to under perform due to the exponential growth in data. Furthermore, to obtain high-performance, these tools require parallel implementations, adding to the development complexity.

Enhancers are short DNA sequences that modulate gene expression patterns. Recent studies have shown that enhancer elements could be enriched for certain histone modification combinatorial codes, leading to interest in developing computational models to predict enhancer locations. Here we present EP-DNN, a protocol for predicting enhancers based on chromatin features, in two different cell types, a human embryonic (H1) and a human lung fibroblast (IMR90) cell line. Specifically, we use a deep neural network (DNN)-based architecture to extract enhancer signatures. We train EP-DNN using distal p300 binding sites, as enhancers, and TSS and random non-DNase-I hypersensitivity sites, as non-enhancers. We find that EP-DNN has superior accuracy relative to other state-of-the-art algorithms, such as DEEP-EN and RFECS, and also scales well to large number of predictions. Then, we surmount the problem that DNN results are not interpretable and develop a method to interpret which histone modifications are important, and within that, which spatial features proximal or distal to the enhancer site, are important. We uncover that the important histone modifications vary between cell types. Further, whether the important features are clustered around the enhancer peak or more spread out also differs among the different histone modifications. Thus, we bring forth a new paradigm for automatically determining the important features and the important histone modifications, rather than the current computational standard of using the same fixed number of features from all the histone modifications for all cell types. Our results have implications for computational scientists who can now do feature selection for their classification task and for biologists who can now experimentally collect data only for the relevant histone modifications.

Gene sequencing instruments are producing huge volumes of data, straining the capabilities of current database searching algorithms and hindering efforts of researchers analyzing large collections of data to obtain greater insights. In the space of parallel genomic sequence search, most of the popular software packages, like mpiBLAST, use the database segmentation approach, wherein the entire database is sharded and searched on different nodes. However this approach does not scale well with the increasing length of individual query sequences as well as the rapid growth in size of sequence databases. In this paper, we propose a fine-grained parallelism technique, called Orion, that divides the input query into an adaptive number of fragments and shards the database. Our technique achieves higher parallelism (and hence speedup) and load balancing than database sharding alone, while maintaining 100% accuracy. We show that it is 12.3X faster than mpiBLAST for solving a relevant comparative genomics problem.