Oases is a de novo RNA-seq assembler that efficiently handles transcriptome assembly across a wide range of expression levels and alternative splicing events. It uses multiple k-mer lengths, dynamic noise filtering, and robust resolution of alternative splicing to merge assemblies. Oases outperforms existing assemblers like transABySS and Trinity, especially in low-expression genes. It is freely available under the GPL license and has been tested on human and mouse RNA-seq data. The software combines multiple k-mer assemblies with topological analysis to produce accurate transcript assemblies. Oases-M, a merged assembly tool, improves results by combining assemblies from different k-mer lengths. It is particularly effective for low-expression genes and handles complex transcript structures. Oases is robust to varying coverage and splicing events, and its performance is validated against reference-based assemblers like Cufflinks. The study shows that Oases provides better overall results compared to previous assemblers, especially in handling low-expression genes and complex splicing events. The software is efficient and scalable, with a runtime and memory usage suitable for large datasets. Overall, Oases is a powerful tool for de novo transcriptome assembly, especially in the absence of a reference genome.Oases is a de novo RNA-seq assembler that efficiently handles transcriptome assembly across a wide range of expression levels and alternative splicing events. It uses multiple k-mer lengths, dynamic noise filtering, and robust resolution of alternative splicing to merge assemblies. Oases outperforms existing assemblers like transABySS and Trinity, especially in low-expression genes. It is freely available under the GPL license and has been tested on human and mouse RNA-seq data. The software combines multiple k-mer assemblies with topological analysis to produce accurate transcript assemblies. Oases-M, a merged assembly tool, improves results by combining assemblies from different k-mer lengths. It is particularly effective for low-expression genes and handles complex transcript structures. Oases is robust to varying coverage and splicing events, and its performance is validated against reference-based assemblers like Cufflinks. The study shows that Oases provides better overall results compared to previous assemblers, especially in handling low-expression genes and complex splicing events. The software is efficient and scalable, with a runtime and memory usage suitable for large datasets. Overall, Oases is a powerful tool for de novo transcriptome assembly, especially in the absence of a reference genome.