Wood is a natural resource that has been an integral part of society for as long as humans have been on Earth. From serving as fuel for fire and construction material the applications of wood has been developed in modern times and only surged as a consequence of societal strive for sustainability and usage of green resources. The idea of disintegrating the cell wall of wood into its cellulosic nanoscale components is one venue that have been explored and garnered intense research activity over the past two decades. The product, referred to as cellulose nanofibrils inherits many of the excellent properties of wood and have subsequently been envisioned in a large variety of applications. When attempting to isolate cellulose nanofibrils from native wood the need for multistep processes are frequently encountered. This reduces the commercial feasibility and makes product characteristics difficult to control. The starting material for academic and commercial endeavors is for this reason often pulp from paper mills. Although less demanding in the form of processing cost, the prospect of controlling and understanding nanofibrils as a function of initial wood properties remains scarcely studied. Subsequently it is difficult to bridge the gap between the properties of wood and the properties of isolated nanofibrils and what processes are required when using wood as a feedstock. This thesis work aims to fill this knowledge gap through i) development of an experimentally robust framework that is feasible for isolation and characterization of nanofibrils from raw wood and ii) implementation to isolate nanofibrils from wood that has been carefully selected through field-grown and genetically engineered aspen trees with varieties in ultrastructure, chemical composition and mechanisms involving cellulose biosynthesis. A one-pot chemical oxidative treatment based on the catalyst 2,2,6,6-tetramethylpiperidine 1-oxyl was adopted, modified and subsequently used for the different wood samples comprising the studies in this work. Wood with a larger cellulose content due to cell wall structural alteration (tension wood) was more difficult to fully disintegrate into fine nanofibrils, and resulted in networks that were twice as tough and comprised of more cellulose with a larger degree of crystallinity and fibril diameter. Wood with a variety in initial lignin content (17 – 30 %) resulted in nanofibrils that were more fibrillated in the case of the highest lignin containing wood despite slightly less degree of oxidation. Estimated surface area of corresponding nanofibrils was higher which implicated wood cell porosity following chemical treatment as a factor of influence on the fibrillation process. Wood from transgenic trees with a reduction in the expression of one of the proteins that is involved for normal cellulose microfibril synthesis (cellulose synthase interacting 1) gave rise to lower aspect ratio nanofibrils with corresponding decreased network toughness and degree of polymerization. Similar characteristics was shown for the initial wood which showcased the possibility to influence nanofibril product quality through genetic engineering of the original tree. The results from this thesis work shows on the possibility to greatly impact nanofibril product quality through the simultaneous design of initial wood properties and appropriate use of processing conditions. This work thus considers a fundamental approach in the sense of having wood as a central feedstock for nanofibrils, something which gives insight in processing and behavior of final nanofibrils and model products. This work opens up for future processing designs which consequently influences further development and final applications of cellulose nanofibrils.