In the absence of a method to ensure that crystals can be obtained for any given protein, the possibility of developing scaffolds for protein crystallisation becomes attractive. Among several approaches that could yield scaffolds, two are particularly promising: the first is based on immunoglobulin Fab fragments and immunoglobulin binding proteins while the second is based on fusion proteins. In the Fab based scaffold, the protein of interest is the antigen recognised by the antibody. In the second case, it is a protein fused to one of the scaffold components. The operational difference between the two methods is the existence of a flexible covalent tether compared to a highly specific interaction. The relative merits and disadvantages of each approach are discussed here. We also describe a lattice obtained through a combinatorial approach which appears to have the required properties to be considered a scaffold. The system makes use of an Fab derived from a rheumatoid factor and an Fc-fusion protein. The Fc-fusion system is ideal for enhanced expression of glycoproteins in mammalian cells and provides a useful tag for their purification. The molecular replacement shows a mode of binding for this rheumatoid factor that is not competitive with bacterial Fc-binding proteins. Hence it may be possible to generalize the method to include bacterial expression of fusion proteins with either protein A or protein G as the fusion partner.