Model transformations occur everywhere in the Model-Driven Development processes. The substrate of these transformations is often an alignment, more or less precise, between a source and a target. There are two main approaches to help the development of model transformations by a semi-automatic generation. The first approach is based on meta-model alignement and searches for mappings that provide transformation rules. In the second one (Model Transformation Based Example), transformation rules are inferred from a set of transformation examples. We develop an integrated MTBE approach where examples are semi-automatically aligned before being used for learning the transformation rules with Formal Concept Analysis.