Executive summary
Within its mission to operate EU surveillance networks, ECDC supports the integration of whole genome sequencing (WGS) data into surveillance and multi-country outbreak investigations of foodborne diseases including listeriosis as one of the priority diseases. To evaluate the inter-laboratory reproducibility and portability of Listeria monocytogenes genome assemblies, ECDC organised a proficiency test for national public health reference laboratories with WGS typing capabilities in the EU/EEA, as well as EFSA and the EU Reference Laboratory for L. monocytogenes.
This report presents the results of the proficiency test. Each participant received a total of 15 sets of raw sequence reads, which were to be assembled by one or more pipelines of their choice. The resulting assemblies were then compared to the reference assembly generated by ECDC on several quality metrics. There were 16 participants, submitting results for 29 pipelines. Twelve participants, including 10 of the 14 participating public health reference laboratories, had at least one concordant pipeline for Illumina reads. The other participants were provided with individual feedback on possibilities to improve their pipeline(s). Participants with a concordant pipeline are recommended to use that for their own analyses as well as for any sharing of assemblies with other organisations including ECDC. For EU-level surveillance purposes ECDC will only accept assemblies generated with a concordant pipeline. Any new pipelines or updates to existing pipelines should go through the same proficiency testing before being used for sharing data with ECDC. For outbreak investigation purposes when more detailed analysis can be needed, raw sequence reads are proposed to be shared instead of or in addition to assemblies for isolates included in the cluster. For Ion Torrent reads, it was not possible to establish concordance. ECDC suggests that any countries producing
these reads share not only the reads with other organisations but also the extracted allele sequences for at least the core genome in the form of a fasta file. This was shown to produce acceptable results and allows other organisations, including ECDC, to perform their allele calling as with any regular assembly. It was also found that the assembly process can be used to remove low-level contamination. Conversely, low-level
contamination can give rise to much longer assembly lengths than the expected length due to the presence of a
large number of very small contigs with very low quality. It is recommended that assembly pipelines include
removal of such small and unreliable contigs, ideally in a way that still alerts the user to the likely presence of
low-level contamination.
EU laboratories that have installed a new or updated pipeline are welcome to have its concordance assessed by
ECDC at any time.
