BUMP: A FORTRAN program for identifying dose-response curves subject to downturns

Douglas G. Simpson, Gerard E. Dallal

Research output: Contribution to journalArticlepeer-review

Abstract

BUMP is a FORTRAN implementation of a modified Jonckheere-Terpstra test, proposed by Simpson and Margolin, to test nonparametrically for a dose-response curve when a downturn is possible at high doses. The Jonckheere-Terpstra statistic is commonly used to test for increasing or decreasing trends in dose-response relationships. In many experimental settings, however, a test agent has more than one effect, and a "bump"-shaped dose-response can occur. For instance, increasing the concentration of a certain nutrient on a petri dish may increase the growth rate at low doses yet decrease the growth rate at high doses because of toxicity. The modified test allows one to assess the significance of the initial increase in the dose-response curve and yet to minimize the effect on the conclusions of any downturn at higher doses. A complete system which operates directly on SYSTAT/MYSTAT files is available for the IBM-PC and compatibles; it includes a utility which converts ASCII data files to the SYSTAT/MYSTAT format. The FORTRAN 77 source code is available for those who would like to run BUMP on other machines.

Original languageEnglish (US)
Pages (from-to)36-43
Number of pages8
JournalComputers and Biomedical Research
Volume22
Issue number1
DOIs
StatePublished - Feb 1989

ASJC Scopus subject areas

  • Medicine (miscellaneous)

Fingerprint Dive into the research topics of 'BUMP: A FORTRAN program for identifying dose-response curves subject to downturns'. Together they form a unique fingerprint.

Cite this