Dataset Information

Towards Program Optimization through Automated Analysis of Numerical Precision.

ABSTRACT: Reducing the arithmetic precision of a computation has real performance implications, including increased speed, decreased power consumption, and a smaller memory footprint. For some architectures, e.g., GPUs, there can be such a large performance difference that using reduced precision is effectively a requirement. The tradeoff is that the accuracy of the computation will be compromised. In this paper we describe a proof assistant and associated static analysis techniques for efficiently bounding numerical and precision-related errors. The programmer/compiler can use these bounds to numerically verify and optimize an application for different input and machine configurations. We present several case study applications that demonstrate the effectiveness of these techniques and the performance benefits that can be achieved with rigorous precision analysis.

SUBMITTER: Linderman MD

PROVIDER: S-EPMC5552069 | biostudies-other | 2010 Apr

REPOSITORIES: biostudies-other

ACCESS DATA

Similar Datasets

Project description:The incidence of new cancer cases is expected to increase significantly in the future, posing a worldwide problem. In this regard, precision oncology and its diagnostic tools are essential for developing personalized cancer treatments. Nowadays, it is almost impossible to separate technology and digitization from medicine and health sciences, and digital pathology (DP) is emerging as one of the most influential technologies in the transition towards the 4P of new medicine (preventive, participatory, personalized and predictive). DP is a particularly key strategy to study the interactions of tumor cells and the tumor microenvironment (TME), which play a crucial role in tumor initiation, progression and metastasis. The purpose of this study was to integrate data on the digital patterns of reticulin fiber scaffolding and the immune cell infiltrate, transcriptomic and epigenetic profiles in aggressive uterine adenocarcinoma (uADC), uterine leiomyosarcoma (uLMS) and their respective lung metastases, with the aim of obtaining key TME biomarkers that can help improve metastatic prediction and shed light on potential therapeutic targets. Automatized algorithms were used to analyze reticulin fiber architecture and immune infiltration in colocalized regions of interest (ROIs) of 133 invasive tumor front (ITF), 89 tumor niches and 70 target tissues in a total of six paired samples of uADC and nine of uLMS. Microdissected tissue from the ITF was employed for transcriptomic and epigenetic studies in primary and metastatic tumors. Reticulin fiber scaffolding was characterized by a large and loose reticular fiber network in uADC, while dense bundles were found in uLMS. Notably, more similarities between reticulin fibers were observed in paired uLMS then paired uADCs. Transcriptomic and multiplex immunofluorescence-based immune profiling showed a higher abundance of T and B cells in primary tumor and in metastatic uADC than uLMS. Moreover, the epigenetic signature of paired samples in uADCs showed more differences than paired samples in uLMS. Some epigenetic variation was also found between the ITF of metastatic uADC and uLMS. Altogether, our data suggest a correlation between morphological and molecular changes at the ITF and the degree of aggressiveness. The use of DP tools for characterizing reticulin scaffolding and immune cell infiltration at the ITF in paired samples together with information provided by omics analyses in a large cohort will hopefully help validate novel biomarkers of tumor aggressiveness, develop new drugs and improve patient quality of life in a much more efficient way.

Dataset Information

Towards Program Optimization through Automated Analysis of Numerical Precision.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets