In the quest for a mathematical measure able to captureand shed light on the dual notions of information and complexity inbiosequences, Hazen et al. have introduced the notion of Functional Information(FI for short). It is also the result of earlier considerationsand findings by Szostak and Carothers et al. Based on the experimentsby Charoters et al., regarding FI in RNA binding activities, we decidedto study the relation existing between FI and classic measures of complexityapplied on protein-DNA interactions on a genome-wide scale. Usingclassic complexity measures, i.e, Shannon entropy and KolmogorovComplexity as both estimated by data compression, we found that FIapplied to protein-DNA interactions is genuinely different from them.Such a fact, together with the non-triviality of the biological functionconsidered, contributes to the establishment of FI as a novel and usefulmeasure of biocomplexity. Remarkably, we also found a relationship, on agenome-wide scale, between the redundancy of a genomic region and itsability to interact with a protein. This latter finding justifies even moresome principles for the design of motif discovery algorithms. Finally,our experiments bring to light methodological limitations of LinguisticComplexity measures, i.e., a class of measures that is a function of thevocabulary richness of a sequence. Indeed, due to the technology and associatedstatistical preprocessing procedures used to conduct our studies,i.e., genome-wide ChIP-chip experiments, that class of measures cannotgive any statistically significant indication about the relation betweencomplexity and function. A serious limitation due to the widespread useof the technology.
|Numero di pagine||13|
|Stato di pubblicazione||Published - 2010|