Reconstructing the insulin secretion rate by Bayesian deconvolution of phase-type densities

  • Andersen, Kim Emil (Projektdeltager)
  • Højbjerre, Malene (Projektdeltager)

Projektdetaljer

Beskrivelse

The rate by which the insulin is secreted from the pancreatic beta-cells is not directly measurable as part of the insulin is absorbed by the liver before entering the blood stream. However, C-peptide is co-secreted equimolarily with the insulin and is not absorbed by the liver, implying that reconstruction of the insulin secretion rate(ISR) can be done by solving a highly ill-posed deconvolution problem for the C-peptide. We represent the ISR, the C-peptide concentration and the convolution kernel as scaled phase-type densities and develop a Bayesian methodology for estimating such densities simultanously via Markov chain Monte Carlo techniques. Hereby closed form evaluation of ISR is possible. We demonstrate the methodology on experimental data from healthy subjects and obtain time-continous ISRs which are more realistic compared to the secretion rates obtained with traditional estimation methods, where the ISR is considered as piecewise constant, and estimated via a two-stage procedure on the basis of an initial estimate of the convolution kernel.
StatusAfsluttet
Effektiv start/slut dato01/09/200631/12/2007

Samarbejdspartnere

  • Vestas (Projektpartner)

Finansiering

  • <ingen navn>

Fingerprint

Udforsk forskningsemnerne, som dette projekt berører. Disse etiketter er oprettet på grundlag af de underliggende bevillinger/legater. Sammen danner de et unikt fingerprint.