Optically exited charge-transfer states play a crucial role in the early stages of photosynthesis but are notoriously difficult to describe efficiently with first-principles computational techniques. This difficulty not only limits the current understanding of the detailed atomistic processes in photosynthesis but is also an obstacle in simulating novel designs of photovoltaic materials that are based on organic absorbers. Current first-principles techniques for electronic excitations suffer either from failure to describe charge-transfer excitations or from poor scalability. This project aims to overcome the limitation in describing such optical excitations by implementing an efficient scheme to solve the Bethe-Salpeter equation, which is the state-of-art method for optical properties of solids and has been shown to correctly describe charge-transfer also in molecules. This method is based on a local-orbital basis for the electronic structure and uses a self-consistent Sternheimer equation to obtain the electronic response aiming to solve the Bethe-Salpeter equation for systems containing several hundreds of atoms and thus considerably broadens the range of what can currently be done with available software. It will be used to study the components of natural light-harvesting complexes that participate in the absorption and charge separation process. The application of this computational method is, however, not limited to these systems. On the contrary, this method, being a first-principles approach, has a wide range of applicability, including technological developments such as the design of photovoltaic devices. The implementation developed in this project will therefore be made available to several widely used electronic structure simulation packages through an interface, so that a wide community of users can profit from this development.