Benchmarking metagenomics classifiers on ancient viral DNA: a simulation study

Owing to technological advances in ancient DNA, it is now possible to sequence viruses from the past to track down their origin and evolution. However, ancient DNA data is considerably more degraded and contaminated than modern data making the identification of ancient viral genomes particularly challenging. Several methods to characterise the modern microbiome (and, within this, the virome) have been developed. Many of them assign sequenced reads to specific taxa to characterise the organisms present in a sample of interest. While these existing tools are routinely used in modern data, their performance when applied to ancient virome data remains unknown. In this work, we conduct an extensive simulation study using public viral sequences to establish which tool is the most suitable for ancient virome studies. We compare the performance of four widely used classifiers, namely Centrifuge, Kraken2, DIAMOND and MetaPhlAn2, in correctly assigning sequencing reads to the corresponding viruses. To do so, we simulate reads by adding noise typical of ancient DNA to a randomly chosen set of publicly available viral sequences and to the human genome. We fragment the DNA into different lengths, add sequencing error and C to T and G to A deamination substitutions at the read termini. Then we measure the resulting precision and sensitivity for all classifiers. Across most simulations, 119 out of the 120 simulated viruses are recovered by Centrifuge, Kraken2 and DIAMOND in contrast to MetaPhlAn2 which recovers only around one third. While deamination damage has little impact on the performance of the classifiers, DIAMOND and Kraken2 cannot classify very short reads. For data with longer fragments, if precision is strongly favoured over sensitivity, DIAMOND performs best. However, since Centrifuge can handle short reads and since it achieves the highest sensitivity and precision at the species level, it is our recommended tool overall. Regardless of the tool used, our simulations indicate that, for ancient human studies, users should use strict filters to remove all reads of potential human origin. Finally, if the goal is to detect a specific virus, given the high variability observed among tested viral sequences, a simulation study to determine if a given tool can recover the virus of interest should be conducted prior to analysing real data.