African swine fever (ASF) is one of the most pathogenic viral diseases of swine leading to clinical and pathomorphological signs of a viral hemorrhagic fever (1). In 2007, ASF was introduced into Georgia (2) and thereafter into numerous eastern European and European Union (EU) countries (3), as well as Asia (China, Mongolia, and Vietnam) (World Organization for Animal Health–World Animal Health Information database, http://www.oie.int/wahis_2/public/wahid.php/Wahidhome/Home/indexcontent, 2019 Feb 21).

In September 2018, the African swine fever virus (ASFV) was introduced into Belgium (3,4); as of April 28, 2019, 719 cases have been reported by Belgium’s Federal Agency for the Safety of the Food Chain (http://www.afsca.be/ppa/actualite/belgique).

Samples from the first 2 cases were taken for analysis to the Belgium National Reference Laboratory for ASF at Sciensano, Brussels, and were later confirmed by the ASF EU Reference Laboratory (EURL) in Valdeolmos, Spain. Partial sequencing at the EURL revealed a p72 genotype II with CVR-1, IGR-2, and MGF1 variants. Initial assessments of virus type and epidemiology were published by Garigliany et al. (4) to share the data without delay. Subsequently, samples were transferred to the Friedrich-Loeffler-Institut, Greifswald, Germany, for whole-genome sequencing.

We prepared samples and sequenced them on an Illumina MiSeq (https://www.illumina.com), as previously described (5). In addition, we enriched DNA libraries for Illumina sequencing for ASFV specific target sequences using an ASFV-specific myBaits kit (Arbor Biosciences, https://arborbiosci.com). We analyzed the resulting sequence data by mapping against an improved ASFV Georgia 2007/1 sequence (International Nucleotide Sequence Database Collaboration accession no. FR682468.2) using Newbler 3.0 software (6); we assembled the identified ASFV-specific reads using SPAdes 3.13.0 (7).

For the assembly of the inverted terminal repeat (ITR) regions, we mapped the reads against the individual ASFV-Georgia 2007/1 ITR regions using Newbler and manually assembled them with the ASFV Belgium 2018/1 sequence in Geneious 11.1.5 (Biomatters, https://www.geneious.com). For validation, we mapped all reads along the final contig using Newbler; the result was a median unique depth of 292 with an interquartile range of 42. We annotated the sequence according to the improved ASFV Georgia 2007/1 sequence, using Glimmer3 in Geneious. We aligned different available ASFV whole-genome sequences using MAFFT 7.388, and we performed a phylogenetic analysis using IQ-TREE v1.6.5 (8,9). The whole-genome sequence is available from the International Nucleotide Sequence Database Collaboration databases under study accession no. PRJEB31287 and sequence accession no. LR536725.

The ASFV Belgium 2018/1 whole-genome sequence has a length of 190,599 bp. Comparison with the improved ASFV Georgia 2007/1 sequence revealed 15 differences, for an overall sequence identity of 99.98% at the nucleotide level. The detected differences included 4 nucleotide transitions, 5 nucleotide transversions, 5 changes in homopolymer regions, and 1 integration of a repeat into a previously described variable intergenic region (10) (Table).

Altogether, 4 differences in annotated genes are nonsynonymous; 2 cause a frameshift, thereby truncating the MGF 110–1L gene and changing the amino acid sequence of the DP60R protein; and 9 differences were identified in noncoding regions (Table). The differences in the ITR regions must be viewed carefully because of the low coverage in these particular parts of the genome, but the differences in the core regions are well supported by the sequencing data (Table). The differences at the specific positions 7,061; 44,586; 134,524; and 170,827 were also identified in the ASFV-SY2018 (China), Estonia 2014, Kashino 04/13 (Russia), and Pol 16/17 (Poland) sequences, and position 170,827 also in ASFV strain Odintsovo_02/14 (Russia). Further genetic differences could be identified in 7 so-called poly G/C regions. Whether these are artifacts originating from sequencing any of the analyzed genomes or pose real differences remains to be determined by the analysis of further sequences from Belgium and other countries, which is in progress.

Finally, the alignment of all publicly available eastern European whole-genome sequences, as well as ASFV-China, shows that all these genomes are nearly identical with identities of more than 99.9% (Appendix Figure).

In conclusion, we provide a whole-genome analysis of ASFV from Belgium, which could show a high overall identity to recent ASFV strains from eastern Europe and China. We also identified locations showing differences from ASFV Georgia 2007/1 in single nucleotides, as well as a previously described repeat insertion. However, because the low mutation rate and the corresponding high genetic stability of the eastern European ASFV strains have hindered the definition of reliable genetic markers thus far, the currently available whole-genome information does not allow for further statements regarding correlations in space and time, and does not provide enough evidence for a more detailed mapping of strain origin.

Although MGF110 was assigned a possible function in preparing the endoplasmic reticulum for viral morphogenesis (11), in the absence of any observations regarding virus attenuation in the field, no conclusion can be drawn on the effect of the observed differences. Therefore, further in vitro and in vivo characterizations using the ASFV Belgium 2018/1 isolate are needed.

In the future, more high-quality whole-genome ASFV sequences might allow identification of genetic markers that could aid high-resolution molecular epidemiology. Coordinated efforts to improve data sharing, together with harmonized protocols under quality assurance, are of utmost importance to interpret results correctly and aid the fight against ASF.

Top