Many scientific disciplines are generating increasingly larger datasets that need analysis to extract meaningful insights. Data Analysis Workflows (DAWs) are often used for orchestrating computational tools to process and analyze these data. As the tools available become more numerous and their combination more complex, manually constructing valid workflows becomes increasingly challenging. On this poster, we focus on the specific challenges of workflows in computational materials science, which often need to be multi-scale workflows that operate across atomic, electronic, and mesoscopic scales. The complexity resulting from this multi-scale nature and different technical implementations requires the collaboration between multiple domain experts. In fact, systematic multiscale modeling has remained an elusive goal in Computational Materials Science for decades. Here, we present our approach to semantically supported composition of workflows in computational materials science as a novel approach towards achieving this goal. Our approach builds upon different frameworks: 1) Pyiron [1], an integrated development environment for computational materials science that can also be used as general purpose workflow manager for high performance computing (HPC) infrastructures, 2) the Semantikon project [2] that develops a semantic layer to formalize scientific knowledge within the Pyiron ecosystem, and 3) the Automated Pipeline Explorer (APE) [3] that automatically composes workflows based on a semantic model and an abstract workflow specification. To illustrate our approach, we demonstrate the automatic composition of a workflow to calculate the crystal vacancy formation energy using APE. This specific workflow exemplifies our approach's ability to manage complexity, which here stems from the interplay of three distinct frameworks rather than the physical concept of a vacancy. The calculation determines the energy required to remove an atom from its regular position in a crystal lattice, creating an empty site known as a vacancy. This physical process is fundamental to understanding a material's diffusion properties, mechanical strength, and behavior at high temperatures. In addition, in order to address a multi-scale problem, we demonstrate a workflow for the calculation of stress-strain curves, which involves the calculation of parameters at atomistic scales (e.g., elastic constants, lattice parameter), as well as the macroscopic description of a real-scale sample. This example demonstrates how semantically supported workflow composition enables a systematic exploration of the various possible workflows, lowering the entry barrier to workflow construction for scientists, enhancing the efficiency and reproducibility of materials research and paving the way to address truly multiscale problems in materials science. Acknowledgement: This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 414984028 – SFB 1404 FONDA. It further benefited from development in the Innovation Platform MaterialsDigital through project funding grant no. 13XP5094E (BAM) and 13XP5094C (MPI SusMat). [1] https://pyiron.org/ [2] https://github.com/pyiron/semantikon[3] https://github.com/sanctuuary/APE