Skip to main content
ARS Home » Plains Area » Clay Center, Nebraska » U.S. Meat Animal Research Center » Livestock Bio-Systems » Research » Publications at this Location » Publication #400594
Research Project: Improving Lifetime Productivity in Swine using Systems Biology and Precision Management Approaches

Location: Livestock Bio-Systems

Title: Application of bayespulse and pulsar otago to quantify luteinizing hormone pulses in ovariectomized gilts

Author
item Lents, Clay
item Wilson, Kyle
item WIJESENA, HIRUNI - Orise Fellow

Submitted to: Molecular Reproduction and Development
Publication Type: Abstract Only
Publication Acceptance Date: 2/10/2023
Publication Date: 8/31/2023
Citation: Lents, C.A., Wilson, K.E., Wijesena, H.R. 2023. Application of bayespulse and pulsar otago to quantify luteinizing hormone pulses in ovariectomized gilts [abstract]. Molecular Reproduction and Development. 90(7):738. https://doi.org/10.1002/mrd.23697.
DOI: https://doi.org/10.1002/mrd.23697

Interpretive Summary:

Technical Abstract: Rapid advances in computer hardware and operating systems have made many standard computer programs for identifying hormone pulses obsolete. The programming language R is one of the most commonly used open-source software for statistical computing with code that can be updated and modified by end-users to create analysis packages. The objective was to evaluate the use of freely available R-based packages for identification and quantification of LH pulses in pigs. BayesPulse (BP) uses a Markov chain Monte Carlo deconvolution algorithm to estimate pulse location and parameters with Bayesian methods. Pulsar Otago (PO) is based on the once widely used PULSAR program. It was adapted to R by retaining all original computational logic except that the smoothing fraction was decoupled from the pattern of located peaks, and the peak splitting was implemented as a finite state rather than nested conditions. The only common output of these programs is pulse frequency. These programs were applied to a dataset in which LH was quantified by radioimmunoassay in blood samples collected every 10-12 min over 6-8 h from OVX gilts (n = 9) and OVX gilts treated with altrenogest (OVX+A; 15 mg/d, for 10 d; n = 10). This data set was chosen because it provided instances of both high and low signal to noise ratio. A general linear ANOVA, fitting treatment and replicate as fixed effects, was used to estimate differences between treatments in LH pulse parameters. Both programs estimated greater (P < 0.05) LH pulse frequency in OVX than OVX+A gilts (0.74 ± 0.10 vs 0.53 ± 0.09; 0.54 ±0.09 vs 0.30 ± 0.08 pulses/h for PO and BP, respectively). BP estimated OVX gilts had lesser (P < 0.05) pulse width and tended (P =0.07) to have lesser pulse mass than OVX+A gilts. Baseline LH and LH pulse half-life estimated from BP did not differ between treatments. PO estimated greater (P < 0.01) mean LH and LH pulse nadir, peak concentrations, and amplitude for OVX than OVX+A gilts. Both programs correctly identified more frequent LH pulses in OVX gilts. BP estimated OVX gilts to have smaller pulses whereas PO indicated pulses were larger in OVX gilts, but these were defined differently in each program. Both programs struggled when signal to noise ratio was low. In this circumstance, BP appeared to resolve location of pulses better, though required negative bias of priors to do so. Results from PO were largely insensitive to changes in program parameters, with the exception of assay variance and smoothing time. Both programs can be used to identify differences in number of LH pulses in pigs. Estimates of pulse size will depend on establishing prior program parameters using data sets of strong signal to noise ratio.