The increasing spread of antibiotic resistance among microorganisms is a dangerous threat for human health. It is therefore necessary to have appropriate tools to understand the phenomenon and monitor its development. Metagenomics has been shown to be a powerful approach for this purpose, with its capability to capture the complexity of the genetic fingerprint of virtually all microorganisms present in an environment, without the bias of the cultivation in laboratory conditions. Pairwise Comparative Modeling (PCM) is a software for prediction of antimicrobial resistance (AMR) from metagenomic samples that has been shown to outperform other tools due to its structure-based alignmentapproach. However, this comes with a higher computational cost. Therefore, the aim of this project was to optimize the allocation of computational resources for the different processes used by the software and to explore different hyperparameters configurations. A new version of the software was developed to enable a more flexible allocation of computational resources and a better portability in different computational environments. Different settings of the hyperparameters were explored and a configuration was found that reduced the CPU hours required to complete the execution of the pipeline by 65.2% compared to default settings, without affecting the sensitivity of the prediction.These results could facilitate the utilization of PCM and its applicability to different contexts, promotingsurveillance programs and research in the field of AMR.