Search for: All records

AlphaFold 3 (AF3) enables atomic-resolution prediction of biomolecular complexes, driving rapidly growing demand across the life sciences. However, its ∼ 750,GB reference database has effectively confined production deployments to systems with shared parallel filesystems, creating a major barrier for scalability. Distributed high-throughput computing (dHTC) platforms offer vast, heterogeneous compute capacity, but fundamentally lack the shared data infrastructure assumed by AF3. We present a data-aware deployment of AF3 for dHTC, implemented on the Center for High Throughput Computing (CHTC) and the Open Science Pool (OSPool). The workflow is decomposed into a CPU-bound data pipeline that executes on nodes with locally staged, scheduler-advertised databases, and a GPU-bound inference pipeline that opportunistically scales across distributed resources. Using CUDA Unified Virtual Memory (UVM), we extend inference beyond physical GPU limits, enabling predictions of ultra-large complexes that exceed device vRAM. By elevating dataset locality to a schedulable resource via HTCondor ClassAds, we eliminate prohibitive per-job data transfers and enable efficient, federated execution. Beyond scaling throughput, we demonstrate that dHTC can support previously infeasible workloads. Together, these results establish dHTC as a viable—and in some regimes superior— execution model for data-intensive structural biology workflows and provide a general blueprint for deploying large, data-intensive applications on distributed cyberinfrastructure. 

Data sharing has become one of the most complex challenges facing researchers in the technological era. Scientific progress increasingly depends on data that must be efficiently stored, managed, processed, and analyzed, placing stringent demands on processing speed and memory capacity. To address these challenges, the Open Science Data Federation (OSDF) was developed. Built on the Open Science Grid (OSG) project’s StashCache and now using the Pelican Platform, OSDF is a data access framework that provides infrastructure and tools for global data access. The Pelican platform introduced several new features and improvements for the OSDF, including a precise GeoIP system, a web interface for each cache and origin, centralized registration, director services, an advertising system to provide information from each host, and more detailed monitoring data for OSDF hosts. Using the Pelican Platform enables a more precise, controllable data-sharing system worldwide. Several benchmarks and tests show that the Pelican Platform can provide a stable, robust system for the Open Science Data Federation. 

Protein language models trained on evolutionary data have emerged as powerful tools for predictive problems involving protein sequence, structure and function. However, these models overlook decades of research into biophysical factors governing protein function. We propose mutational effect transfer learning (METL), a protein language model framework that unites advanced machine learning and biophysical modeling. Using the METL framework, we pretrain transformer-based neural networks on biophysical simulation data to capture fundamental relationships between protein sequence, structure and energetics. We fine-tune METL on experimental sequence–function data to harness these biophysical signals and apply them when predicting protein properties like thermostability, catalytic activity and fluorescence. METL excels in challenging protein engineering tasks like generalizing from small training sets and position extrapolation, although existing methods that train on evolutionary signals remain powerful for many types of experimental assays. We demonstrate METL’s ability to design functional green fluorescent protein variants when trained on only 64 examples, showcasing the potential of biophysics-based protein language models for protein engineering. 

Context. In a series of publications, we describe a comprehensive comparison of Event Horizon Telescope (EHT) data with theoretical models of the observed Sagittarius A* (Sgr A^*) and Messier 87* (M87^*) horizon-scale sources. Aims. In this article, we report on improvements made to our observational data reduction pipeline and present the generation of observables derived from the EHT models. We make use of ray-traced general relativistic magnetohydrodynamic simulations that are based on different black hole spacetime metrics and accretion physics parameters. These broad classes of models provide a good representation of the primary targets observed by the EHT. Methods. We describe how we combined multiple frequency bands and polarization channels of the observational data to improve our fringe-finding sensitivity and stabilization of atmospheric phase fluctuations. To generate realistic synthetic data from our models, we took the signal path as well as the calibration process, and thereby the aforementioned improvements, into account. We could thus produce synthetic visibilities akin to calibrated EHT data and identify salient features for the discrimination of model parameters. Results. We have produced a library consisting of an unparalleled 962 000 synthetic Sgr A^*and M87^*datasets. In terms of baseline coverage and noise properties, the library encompasses 2017 EHT measurements as well as future observations with an extended telescope array. Conclusions. We differentiate between robust visibility data products related to model features and data products that are strongly affected by data corruption effects. Parameter inference is mostly limited by intrinsic model variability, which highlights the importance of long-term monitoring observations with the EHT. In later papers in this series, we will show how a Bayesian neural network trained on our synthetic data is capable of dealing with the model variability and extracting physical parameters from EHT observations. With our calibration improvements, our newly reduced EHT datasets have a considerably better quality compared to previously analyzed data. 

Due to the increased demand of network traffic expected during the HL-LHC era, the T2 sites in the USA will be required to have 400Gbps of available bandwidth to their storage solution. With the above in mind we are pursuing a scale test of XRootD software when used to perform Third Party Copy transfers using the HTTP protocol. Our main objective is to understand the possible limitations in the software stack to achieve the target transfer rate; to that end we have set up a testbed of multiple XRootD servers in both UCSD and Caltech which are connected through a dedicated link capable of 400 Gbps end-to-end. Building upon our experience deploying containerized XRootD servers, we use Kubernetes to easily deploy and test different configurations of our testbed. In this work, we will present our experience doing these tests and the lessons learned. 

The OSG-operated Open Science Pool is an HTCondor-based virtual cluster that aggregates resources from compute clusters provided by several organizations. Most of the resources are not owned by OSG, so demand-based dynamic provisioning is important for maximizing usage without incurring excessive waste. OSG has long relied on GlideinWMS for most of its resource provisioning needs but is limited to resources that provide a Grid-compliant Compute Entrypoint. To work around this limitation, the OSG Software Team has developed a glidein container that resource providers could use to directly contribute to the OSPool. The problem with that approach is that it is not demand-driven, relegating it to backfill scenarios only. To address this limitation, a demand-driven direct provisioner of Kubernetes resources has been developed and successfully used on the NRP. The setup still relies on the OSG-maintained backfill container image but automates the provisioning matchmaking and successive requests. That provisioner has also been extended to support Lancium, a green computing cloud provider with a Kubernetes-like proprietary interface. The provisioner logic has been intentionally kept very simple, making this extension a low-cost project. Both NRP and Lancium resources have been provisioned exclusively using this mechanism for many months. 

Creating new materials, discovering new drugs, and simulating systems are essential processes for research and innovation and require substantial computational power. While many applications can be split into many smaller independent tasks, some cannot and may take hours or weeks to run to completion. To better manage those longer-running jobs, it would be desirable to stop them at any arbitrary point in time and later continue their computation on another compute resource; this is usually referred to as checkpointing. While some applications can manage checkpointing programmatically, it would be preferable if the batch scheduling system could do that independently. This paper evaluates the feasibility of using CRIU (Checkpoint Restore in Userspace), an open-source tool for the GNU/Linux environments, emphasizing the OSG’s OSPool HTCondor setup. CRIU allows checkpointing the process state into a disk image and can deal with both open files and established network connections seamlessly. Furthermore, it can checkpoint traditional Linux processes and containerized workloads. The functionality seems adequate for many scenarios supported in the OSPool. However, some limitations prevent it from being usable in all circumstances.