incomputable complexity, it is actually amanoeuverthat allows
biologists to unveil the properties of specific objects that are essen-
tial in order to understand the behavior of bigger systems. This
ontological assumption also implies that there is a privileged level of
organization that is causally responsible and explanatorily relevant
for all the other levels. If such level of organization is responsible for
the others, the direction of causation will be from this level to the
others; therefore, offering an explanation at that level should, in
principle, answer all questions arising at the other different levels.
Sometimes, this is labeled the thesis ofbottom-up causation, mean-
ing that lower-levels cause the existence of higher-levels. In molec-
ular biology, the privileged level is the one of macromolecules,
mainly genes or their direct functional products (i.e., proteins).
Let us consider a concrete example. In cancer studies,the
somatic mutation theory(SMT) exemplifies what we have men-
tioned above. In a nutshell, the SMT is the view that cancer is a
disease related to some specific genes. Whatever we observe about
high-level dynamics in cancer, it is caused by mutations of nucleo-
tides at the genomic level. The origin of the view that the genome
sequence is fundamental to explain mostly everything we observe
about biology is controversial and we do not want to discuss it here.
It suffices to say that, in the SMT view, somatic mutations provoke
some sort of advantage to a cell, such that it proliferates faster or
becomes virtually immortal. Therefore, cancer is a complex phe-
nomenon caused by mutations of nucleotides. The level of organi-
zation where to look to explain cancer isthe genome. For this
reason, well-funded research projects today are focused on the
identification of mutations—and in general structural variations—
in the genome (seerefs. 7, 8), because mutations are what causes
cancer. Any (mechanistic) explanation of cancer will then be
focused on the study of how mutated genes do what they do.3 A Relational Ontology for Systems Biology
Several studies, especially cancer studies, show that the derived
SMT intrinsic properties-based, causally-linear ontology is prob-
lematic [9]. The stochastic evolution of cancer, by definition, makes
it impossible to establish direct causal relationships with specific
genetic or epigenetic features. Reconstructing discrete stages is
difficult, and attributing the origin of cancer to a unique intracellu-
lar molecular component or specific exogenous factor seems impos-
sible. During the neoplastic process the molecular components are
mainly unvaried, but their functional activity changes, due to inter-
nal and external factors that eventually involve multiple
DNA-damaging events as well. Such change is considered
dis-functional as far as it does not respond to the normal regulative
factors properly (e.g., aberrant differentiation) and brings about aConceptual Challenges in Systems Biology 3