When the DNA is too degraded, as is often the case with ancient DNA, or when a quick identification is required that focuses on repeated zones, or when conventional mapping tools prove inadequate, this detection pipeline identifies species by comparing k-mers dictionaries between a chosen reference genome and the studied samples requiring taxonomic determination. A k-mer is a subsequence of length k nucleotides extracted from a longer sequence. This pipeline includes an optional sample fastp cleaning step before manipulating k-mers with KMC_tools [1] to generate k-mers dictionaries for one or more reference genomes and for samples studied, and intersect (pairwise comparisons) these dictionaries. Then a python script calculates several indicators like the Jaccard similarity indices [2] in order to identify sample to a species. This pipeline allows for executing the dictionary comparison step as many times as needed to compare the sample to a given reference genome. By comparing indices, the pipeline determines the species closest to studied sample. Afterward, the results are graphically displayed in an HTML report file which containing Hightcharts graphics. This pipeline is being tested on both ancient and modern DNA, on several animal species and on genomes of different sizes. Indeed, we are studying the relationship between sequence quality and the potential impact on the Jaccard similarity index obtained. To minimize dictionaries sizes, either filter out low-quality k-mers and erroneous k-mers within the expected sequencing depth or work only on repeated zones. If working on the k-mers of the repeated zones yields favorable outcomes, it may suggest that these repeated zones are species-specific and provide an indication of their resistance to DNA degradation. This potential impact of repeated zones is being studied by comparing small dictionaries, limited to 1000 k-mers, representing those most commonly found within a species. This may indicate whether the repeats are genome-specific, thereby serving as identifiers for the genome. Working only the genome-specific repetitive would save time to generate k-mers analysis sets. For some species, we intend to provide biologists with precomputed sets of k-mers to optimize processing times and computing resources. This pipeline has previously been employed to detect fish species from ancient DNA [3]. The pipeline name is DeFIS « Detection of Fauna and flora: Identification of Species ». Acknowledgment to partners who provided samples of ancient and modern DNA on which we tested this computational tool: Barbaza Michel, Clavel Benoît, Dachary Morgane, Daujeard Camille, Gey Delphine, Guillaud Émilie, Hamilton-Dyer Sheila, Harland Jennifer, Hinguant Stéphan, Horad-Herbin Marie-Pierre, Hummel Esmee, Van Der Jagt Inge, Jonsson Leif, Madelaine Stéphane, Makowiecki Daniel, Perrin Thomas, Primault Jérôme, Quinlan Liz, Schmölcke Ulrich