**Background:** The white spot syndrome virus (WSSV) is the largest known animal virus, with a genome ranging from 280–314 Kb and over 500 coding sequences (CDS). It causes devastating economic losses in global shrimp aquaculture, which produced 6.55 million tonnes valued at USD 40.67 billion in 2019. Since its first report in the early 1990s, WSSV has caused billions of USD in losses and can induce 100% cumulative mortality within 3 days. A critical barrier to WSSV research is inconsistent genome annotation nomenclature across isolates, arising from its novel genome sequence, circular DNA, variable genome length, and genome-wide deletions. Different isolates use different prefixes (e.g., wsv001, WSSV218, ORF90) and numbering systems, making cross-isolate comparisons nearly impossible. The authors developed Missing Regions Finder (MRF) to address this by documenting missing genomic regions and CDS in one uniform annotation nomenclature.
**Methods:** MRF combines the MUMmer tool (with the 'mum' option for exact base matches) with custom Perl scripts (v5.20.2) and Bash (v4.3.30). It requires three inputs: a query genome sequence, a reference genome sequence, and its annotation table in GFF3 format. MRF calculates exact matching genome regions between query and reference, then identifies missing regions (MR) between consecutive matches. It compares MR coordinates with CDS positions in the GFF3 table to output completely and partially missing CDS. The tool is available as a web interface (Bootstrap 4, PHP 5.6.33) and a command-line utility (single and batch modes). Advanced features include adjustable exact match length (-m, default 20), false match length cutoff (-l, default 15), and negative/positive offset parameters (-n/-p, default 1/1) to filter false matches. Outputs include tables of missing CDS, missing genomic regions, completely present CDS, and a MirCos plot (Circos-based) depicting missing CDS. For WSSV comparative genomics, 13 complete genomes (280,591–309,286 bp, GC content 40.85–41.08%) from 8 countries were analyzed against the CN isolate (AF332093) as reference. Phylogenomic analyses used MAFFT v7.305b (FFT-NS-2 algorithm), RAxML v8.2.9 (GTRGAMMAI, 1000 bootstrap replicates), and MrBayes 3.2.6 (two independent runs, 4 chains, 10 million generations, 25% burn-in). A random forest model (Python scikit-learn v0.21.1) ranked important CDS for isolate clustering. Homology searches used Blast2GO v4.1.9 (blastn, e-value 1e-03, alignment length >200 bp) against Penaeus vannamei, Eriocheir sinensis, and Homo sapiens genomes.
**Key Results:** Compared to the CN isolate (524 CDS), 82 CDS were completely or partially lost in at least one other isolate, with 46 CDS showing complete deletion in at least one isolate. Three deletion hotspots were identified: wsv481/wsv499 (lost in 6 isolates), wsv237/wsv241 (lost in 6 isolates), and wsv178/wsv180 (lost in 3 isolates). The random forest model ranked wsv237, wsv485, and wsv481 as the top three CDS important for isolate clustering. Phylogenetic analysis revealed three clusters with a potential correlation to virulence (CN01 = high-virulent, CN02 = medium-virulent, CN03 = low-virulent). Essential CDS retained across all isolates included structural proteins VP19, VP26, VP28; all four per os infectivity factors (PIFs: wsv35, wsv115, wsv209, wsv306); four RING-H2 domain proteins (wsv199, wsv222, wsv249, wsv403); VP9 (wsv230); and five nucleotide metabolism proteins (wsv67, wsv112, wsv172, wsv188, wsv395). Only 4 of 524 WSSV CDS showed significant homology to shrimp genes: wsv67 (thymidylate synthetase, 553 bp), wsv91 (immediate early protein, 240 bp), wsv172 (ribonucleotide reductase large subunit, 545 bp), and wsv285 (202 bp). MRF demonstrated advantages over BLAST in handling low-complexity regions (wsv095, wsv164), repeat-rich regions (wsv178, where BLAST reported 100% query coverage but MRF detected 58.09% deletion), and highly similar gene families (ASFV MGF 505-2R, where BLAST produced a false alignment with MGF 505-10R). MRF also successfully detected an 18 bp deletion in MDV UL49 gene and tracked ~400 bp deletions in HIV-1 Nef gene.
**Clinical Implications:** MRF provides a unique tool for comparative genomics of large, highly similar, length-varying, and inconsistently annotated viral genomes by focusing on what is absent rather than what is present. For WSSV, the tool enables rapid identification of CDS deletions across isolates in a single annotation nomenclature, facilitating correlation of genomic differences with virulence. The finding that ubiquitination, transcription regulation, and nucleotide metabolism proteins are retained across all isolates suggests these functions are essential for WSSV pathogenesis. The identification of 14 CDS with predicted glycosylation sites (including 6 with GO terms for viral nucleocapsid, envelope glycoprotein, or membrane components) that are deleted in medium- and low-virulent isolates suggests these minor structural proteins may function as envelope glycoproteins for cell attachment. MRF's batch mode allows researchers to quickly identify the most important CDS and genomes requiring focused attention from large datasets. The tool is freely available at https://github.com/vinayciba/MRF and http://14.139.181.163/mrf.