!"#$#%&’ )&*+" ,11. Matthäus F et al. Interactive Molecular Networks … Pharmacopsychiatry 2009; 42 (Suppl. 1): S118–S128 Bibliography DOI 10.1055/s-0029-1216348 Pharmacopsychiatry 2009; 42(Suppl. 1): S118–S128 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0936-9528 Correspondence Prof. Dr. P. J. Gebicke-Haerter Department of Psychopharmacology Central Institute for Mental Health J5 68159 Mannheim Tel: + 49/621/170 362 56 Fax: + 49/621/170 362 55 peter.gebicke@zi-mannheim.de Interactive Molecular Networks Obtained by Computer-aided Conversion of Microarray Data from Brains of Alcohol-drinking Rats play today an important role in obtaining expres- sion data of many genes measured simultane- ously. After ample experience with some generations of expression proling platforms from deRisi&Brown and A!ymetrix to Agilent’s and Illumina’s [7, 8, 37], we are facing substantial improvements both in terms of tissue prepara- tion and reliability and precision of microarrays. This has been accompanied by increasing num- bers of genes with statistically signicant p-val- ues identied as di!erentially expressed in those conditions. Whilst only a few dozens of genes ful- lled those criteria previously, very often several hundreds are detected more recently. Because also statistical tools of processing chip data underwent considerable amendments, occur- rences of false positives or false negatives in those datasets have been reduced as well. This increased trustworthiness into the data has come along with a great deal of confusion to understand and Introduction & In recent years, moving together with techno- logical progress, modern molecular biology has evolved from focusing on single cell components to the analysis of whole biological systems. Now- adays scientists are able to perform global quali- tative and quantitative analysis of whole networks of molecular interactions within a cell. This has spurred scientists to generate a new branch of biological sciences, whose aim is to understand the functional aspects of entire sys- tems, namely, systems biology. An important task here is to explore the eld of gene and gene prod- uct relationships. Understanding mechanisms and dependencies within gene regulatory net- works (GRNs) is crucial for obtaining more detailed insights into pathological processes, and for further drug target identication. DNA micro- arrays, which evolved in the middle 1 990 s [45], Authors F. Matthäus 1 , V.A. Smith 2 , A. Fogtman 3, 4 , W.H. Sommer 5 , F. Leonardi-Essmann 5 , A. Lourdusamy 6 , M.A. Reimers 7 , R. Spanagel 5 , P.J. Gebicke-Haerter 5 A!liations A!liation addresses are listed at the end of the article Abstract & Lists of di!erentially expressed genes in a disease have become increasingly more comprehen- sive with improvements on all technical levels. Despite statistical cuto!s of 99 % or 95 % con- dence intervals, the number of genes can rise to several hundreds or even thousands, which is barely amenable to a researcher’s understand- ing. This report describes some ways of process- ing those data by mathematical algorithms. Gene lists obtained from 53 microarrays (two brain regions (amygdala and caudate putamen), three rat strains drinking alcohol or being abstinent) have been used. They resulted from analyses on A!ymetrix chips and encompassed approxi- mately 6 000 genes that passed our quality lters. They have been subjected to four mathematical ways of processing: (a) basic statistics, (b) princi- pal component analysis, (c) hierarchical cluster- ing, and (d) introduction into Bayesian networks. It turns out, by using the p-values or the log- ratios, that they best subdivide into brain areas, followed by a fairly good discrimination into the rat strains and the least good discrimination into alcohol-drinking vs. abstinent. Neverthe- less, despite the fact that the relation to alco- hol-drinking was the weakest signal, attempts have been made to integrate the genes related to alcohol-drinking into Bayesian networks to learn more about their inter-relationships. The study shows, that the tools employed here are extremely useful for (a) quality control of data- sets, (b) for constructing interactive (molecular) networks, but (c) have limitations in integration of larger numbers into the networks. The study also shows that it is often pivotal to balance out the number of experimental conditions with the number of animals.