Fungal infections in humans affect 25% (~1.7 billion) of the world population, with invasive fungal disease (IFD) alone causing 1.6 million deaths/year. They are treatable and result in good outcomes if diagnosed in a timely manner. However, current mycological diagnostics take days to weeks, lack specificity and sensitivity, leading to vital delays in treatment, inappropriate therapy, increasing morbidity and mortality and high healthcare costs. Therefore, early identification of the causative pathogen is vital to improve disease outcome. New technologies, such as long-read sequencing-based metagenomics, having the potential to radically transform clinical diagnosis, as they permit real-time detection of multiple pathogens directly from clinical specimens (metabarcoding). The utilization of these technologies is currently hampered due to the lack of standardized protocols, low pathogen:host DNA ratio, sub-optimal DNA quality, limited reference data and inadequate bioinformatics algorithms. To evaluate and compare the resolution of amplicon based short-read Illumina and metagenomics based long-read MinIONsequencing, first a pilot study was conducted with a defined “mock” community of 12 fungal species. Illumina correctly identified 7/12 and MinION6/12 species. Second, metagenomics based long-read MinION sequencing was applied to detect Pneumocystis jirovecii, the non-culturable agent of Pneumocystispneumonia, directly from BAL and sputum specimens. Of the total reads obtained 70 - 95% were assigned a human origin, while most microbial reads were classified as bacteria and only ~10% were fungi.Comparison of all fungal and bacterial reads obtained by WIMP and BLAST for species concordance showed a good correlation for bacterial reads (70.2-97.4%) but not for fungal reads (0.19-10.2%). Whilst P. jiroveciiwas identified, it did not produce the strongest signal, indicating that patient fungal loads, false-positives and sequencing errors need to be taken into account in analysis algorithms. Long-read based metagenomics as diagnostic tool for fungal infections will result in a drastic reduction in turn-around time to 24-48h and significant improvement of accuracy, leading to a more cost-effective and highly improved patient care.