Proteinortho is a tool for detecting (co-)orthologs in large-scale genomic data. It uses a modified reciprocal best alignment heuristic and is optimized for large datasets, reducing memory requirements compared to existing tools. The software is designed for multi-core hardware and can be run on off-the-shelf computers. It processes blast alignments on the fly, avoiding the need to store all pairwise comparisons. Proteinortho was tested on 717 bacterial genomes and identified 30 proteins present in 99% of all bacterial proteomes. The tool uses spectral partitioning to identify orthologous groups, which are dense subgraphs in a graph representation of protein similarities. It outperforms other tools like OrthoMCL in terms of speed and memory usage, especially for large datasets. Proteinortho is suitable for large-scale analysis and can be used to identify conserved proteins across bacterial species, which are useful for phylogenetic studies and as markers for novel bacterial species. The software is available under the GPLv2 license and has been validated against existing databases like the COG database.Proteinortho is a tool for detecting (co-)orthologs in large-scale genomic data. It uses a modified reciprocal best alignment heuristic and is optimized for large datasets, reducing memory requirements compared to existing tools. The software is designed for multi-core hardware and can be run on off-the-shelf computers. It processes blast alignments on the fly, avoiding the need to store all pairwise comparisons. Proteinortho was tested on 717 bacterial genomes and identified 30 proteins present in 99% of all bacterial proteomes. The tool uses spectral partitioning to identify orthologous groups, which are dense subgraphs in a graph representation of protein similarities. It outperforms other tools like OrthoMCL in terms of speed and memory usage, especially for large datasets. Proteinortho is suitable for large-scale analysis and can be used to identify conserved proteins across bacterial species, which are useful for phylogenetic studies and as markers for novel bacterial species. The software is available under the GPLv2 license and has been validated against existing databases like the COG database.