Genomic analyses have revealed that free-living biological organisms carry between 107 and 1011 bits of information in their genomes. In large organisms with relatively small population sizes, such as humans, only in the order of 1% of the genomic information is shaped by the environment via natural selection. A much larger amount of information than this is routinely being generated by biomedical researchers, and the rapidly accumulating data is often interpreted to mean that biological systems are extremely complex. However, as the genome is finite in length, it cannot define precisely optimal values for the quantitative parameters of the experimentally identified molecular phenotypes. Furthermore, because the genomic sequences orchestrate a biochemical system that is much more information-rich than the genome, the vast majority of the measured molecular phenotypes must represent "molecular spandrels", that is phenotypes that are not independent of each other, and instead co-determined by the same genomic sequences. These considerations are important in interpreting the results of individual experiments. In addition, they indicate that full understanding of biological systems requires a genome-centric model that does not abstract away the information contained in the genome, and instead explicitly maps all phenotypic data back to specific genomic sequences.
Original Article: Informational limits of biological organisms