Dataset Information

Predicting runtimes of bioinformatics tools based on historical data: five years of Galaxy usage.

ABSTRACT: MOTIVATION:One of the many technical challenges that arises when scheduling bioinformatics analyses at scale is determining the appropriate amount of memory and processing resources. Both over- and under-allocation leads to an inefficient use of computational infrastructure. Over allocation locks resources that could otherwise be used for other analyses. Under-allocation causes job failure and requires analyses to be repeated with a larger memory or runtime allowance. We address this challenge by using a historical dataset of bioinformatics analyses run on the Galaxy platform to demonstrate the feasibility of an online service for resource requirement estimation. RESULTS:Here we introduced the Galaxy job run dataset and tested popular machine learning models on the task of resource usage prediction. We include three popular forest models: the extra trees regressor, the gradient boosting regressor and the random forest regressor, and find that random forests perform best in the runtime prediction task. We also present two methods of choosing walltimes for previously unseen jobs. Quantile regression forests are more accurate in their predictions, and grant the ability to improve performance by changing the confidence of the estimates. However, the sizes of the confidence intervals are variable and cannot be absolutely constrained. Random forest classifiers address this problem by providing control over the size of the prediction intervals with an accuracy that is comparable to that of the regressor. We show that estimating the memory requirements of a job is possible using the same methods, which as far as we know, has not been done before. Such estimation can be highly beneficial for accurate resource allocation. AVAILABILITY AND IMPLEMENTATION:Source code available at https://github.com/atyryshkina/algorithm-performance-analysis, implemented in Python. SUPPLEMENTARY INFORMATION:Supplementary data are available at Bioinformatics online.

SUBMITTER: Tyryshkina A

PROVIDER: S-EPMC6931352 | biostudies-literature | 2019 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:We examined the adequacy of existing shielding guidelines using five-year clinical data from a busy CyberKnife center. From June 2006 through July 2011, 1,370 patients were treated with a total of 4,900 fractions and 680,691 radiation beams using a G4 CyberKnife. Prescription dose and total monitor units (MU) were analyzed to estimate the shielding workload and modulation factor. In addition, based on the beam's radiation source position, targeting position, MU, and beam collimator size, the MATLAB program was used to project each beam toward the shielding barrier. The summation of the projections evaluates the distribution of the shielding load. On average, each patient received 3.6 fractions, with an average 9.1 Gy per fraction prescribed at the 71.1% isodose line, using 133.7 beams and 6,200 MU. Intracranial patients received an average of 2.7 fractions, with 8.6 Gy per fraction prescribed at the 71.4% isodose line, using 133 beams and 5,083 MU. Extracranial patients received an average of 3.94 fractions, with 9.2 Gy per frac- tion prescribed at the 71% isodose line, using 134 beams and 6,514 MU. Most- used collimator sizes for intracranial patients were smaller (7.5 to 20 mm) than for extracranial patients (20 to 40 mm). Eighty-five percent of the beams exited through the floor, and about 40% of the surrounding wall area received no direct beam. For the rest of the wall, we found "hot" areas that received above-average MU. The locations of these areas were correlated with the projection of the nodes for extracranial treatments. In comparison, the beam projections on the wall were more spread for intracranial treatments. The maximum MU any area received from intracranial treatment was less than 0.25% of total MU used for intracranial treatments, and was less than 1.2% of total MU used for extracranial treatments. The combination of workload, modulation factor, and use factor in our practice are about tenfold less than recommendations in the existing CyberKnife shielding guidelines. The current guidelines were found to be adequate for shielding, even in a busy center. There may be a potential to reduce shielding in areas with no or few direct beams in the current G4 model. Since a newer model CyberKnife (M6) has recently been introduced, the patterns of usage reported here may be changed in the future. The uneven distribution of use factor we have found may, however, be considered in the vault design.

Dataset Information

Predicting runtimes of bioinformatics tools based on historical data: five years of Galaxy usage.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets