Factorial graphical models have recently been proposed for inferring dynamic regulatory networks from high-throughput data. In the search of true regulatory relationships amongst the vast space of possible networks, these models allow to impose certain restrictions on the dynamic nature of these relationships, such as that Markov dependencies are of low order, i.e. some entries of the precision matrix are a priori zeros, or that the strength of the dependencies depend only on time lags, i.e. some entries of the precision matrix are assumed to be equal. The precision matrix is then estimated by $l_1$ penalised likelihood, imposing a further constraint on the absolute value of its entries, which results in sparse networks. The problem of selecting the optimal sparsity level is traditionally framed in terms of the Kullback-Leibler (KL) divergence. In this paper, we present a KL-motivated model selection criterion for factorial graphical models, by taking into account the a priori structural constraints. We test the performance of this method on simulated data and compare it with existing approaches. Finally, we present an application on a detailed time-course microarray data from the \textit{Neisseria meningitidis} bacterium, a causative agent of life-threatening infections such as meningitis.

Vinciotti, V., Augugliaro, L., Abbruzzo, A., Saunders, N., Wit, E. (2014). Model selection for structured dynamic gene regulatory networks of Neisseria meningitis. STATISTICAL APPLICATIONS IN GENETICS AND MOLECULAR BIOLOGY, 00.

Model selection for structured dynamic gene regulatory networks of Neisseria meningitis

2014-01-01

Abstract

Factorial graphical models have recently been proposed for inferring dynamic regulatory networks from high-throughput data. In the search of true regulatory relationships amongst the vast space of possible networks, these models allow to impose certain restrictions on the dynamic nature of these relationships, such as that Markov dependencies are of low order, i.e. some entries of the precision matrix are a priori zeros, or that the strength of the dependencies depend only on time lags, i.e. some entries of the precision matrix are assumed to be equal. The precision matrix is then estimated by $l_1$ penalised likelihood, imposing a further constraint on the absolute value of its entries, which results in sparse networks. The problem of selecting the optimal sparsity level is traditionally framed in terms of the Kullback-Leibler (KL) divergence. In this paper, we present a KL-motivated model selection criterion for factorial graphical models, by taking into account the a priori structural constraints. We test the performance of this method on simulated data and compare it with existing approaches. Finally, we present an application on a detailed time-course microarray data from the \textit{Neisseria meningitidis} bacterium, a causative agent of life-threatening infections such as meningitis.
2014
Settore SECS-S/01 - Statistica
Vinciotti, V., Augugliaro, L., Abbruzzo, A., Saunders, N., Wit, E. (2014). Model selection for structured dynamic gene regulatory networks of Neisseria meningitis. STATISTICAL APPLICATIONS IN GENETICS AND MOLECULAR BIOLOGY, 00.
File in questo prodotto:
File Dimensione Formato  
12_articolo.pdf

accesso aperto

Dimensione 245.23 kB
Formato Adobe PDF
245.23 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10447/101820
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact