RotaC: A web-based tool for the complete genome classification of group A rotaviruses

Group A rotaviruses are the major etiological agent of severe diarrhea in infants and young children worldwide, leading to significant morbidity and mortality. More than 125 million infants and young children develop rotavirus diarrhea globally each year, resulting in 440.000 deaths in children, mostly in the developing countries [1]. Although the infant mortality rate due to rotavirus disease is low in developed countries, the consequences of the disease can be very costly and cause a significant economic burden, which can be both direct (medical costs, outpatient visits, diagnosis, medication) and indirect (lost working hours of parents). For example, the costs associated with rotavirus diarrhea in the United States were estimated at $100-400 million to the healthcare system and $1 billion to the society [2, 3].

The classification tool for group A rotaviruses (RotaC v1.0) is written in java with a simple object model in order to make it easy to maintain the code. The interface of the website is written in perl. The RotaC tool can analyze up to a 1000 nucleotide sequences in 'strict' FASTA-format (a first line with a sequence identifier preceded by '>', followed by a second line with the sequence). The analysis of nucleotide sequences with a length below 500 bases is not suitable according to the RCWG guidelines and is not allowed in the RotaC tool.

The genotyping process consists of several subsequent steps. In a first step, the appropriate gene segment is identified by comparing the query sequence with a full genome reference alignment consisting of well-characterized group A rotavirus sequences and by the neighbor-joining algorithm. After the recognition of the segment of origin, the query sequence is aligned using the profile alignment functions of Clustal W v2.0[7] with a reference alignment of the appropriate segment (detailed information about the alignments used with the RotaC tool can be found on http://rotac.regatools.be). In a second step, a distance matrix, based on pairwise alignments with the Needleman-Wunsch algorithm [8], and a phylogenetic tree based on the neighbor-joining algorithm using the Paup* software [9] are constructed and analyzed to identify the genotype of the query sequence by using the nucleotide identity cut-off values summarized in Table 1. The reliability of the clustering of the neighbor-joining tree is assessed using 100 bootstrap replicates, considering 70% as the cut-off value. If the query sequence has a shared identity of at least 3% above the appropriate cut-off value with an established genotype, the query sequence is considered as a member of that specific genotype. If the shared identity is at least 3% below the cut-off value, the query sequence is considered as a new genotype of the proper rotavirus segment. For identities less than 3% below or above the cut-off value, the tool provides only tentative conclusions. In this case, it is recommended to send the sequence to the Rotavirus Classification Working Group for further phylogenetic analysis and correct identification of the genotype. For queries covering less than 50% of the ORF region, no conclusion will be drawn. The user should pay attention that the assignment of a genotype to the query sequence with identities less than 3% below or above the genotype cut-off values should be confirmed by more extensive phylogenetic analysis, or should be send to the RCWG for further analyses and/or validation.

Due to the very limited number of completely sequenced rotavirus genomes, studies on reassortments have been limited to a few gene segments. Recently, the increased availability of complete rotavirus genome sequences, and the introduction of an extended classification and nomenclature system, comprising all 11 rotavirus gene segments, has prompted many investigators to start complete rotavirus genome sequencing projects. Both reassortments between strains belonging to the same host species, and between strains belonging to different host species have been documented several times in the past [10–12]. The new classification system creates a necessary framework to thoroughly analyze possible interspecies transmissions of whole rotaviruses from one host to another, and to study the effect of reassortments on the generation of genetic rotavirus diversity, host range restriction, co-segregation of certain gene segments, and adaptation to a new host species [5]. A Rotavirus Classification Work Group was setup to evaluate potentially new genotypes that will be discovered when more and more complete rotavirus genomes from multiple host species will be sequenced [6]. The analyses of complete rotavirus genomes, and the assignment to the appropriate genotypes will be highly facilitated by the use of the free online RotaC-tool. The RotaC-tool will be updated regularly, an will work closely together with the RCWG in order to update the tool with new genotypes, to reflect all established and new genotypes.

There are several useful web-based tools and database resources for the genotyping analysis of viral sequences, based on phylogenetic trees, or sequence similarities of whole/partial sequences for the genotyping of HIV-1/HIV-2, HTLV-1/HTLV-2, hepatitis B virus, hepatitis C virus and poliovirus sequences [13–17]. Here we have introduced a reliable and easy-to-use automated classification tool for group A rotaviruses. Our RotaC classification tool is in agreement with the rotavirus classification strategy and guidelines as proposed by the Rotavirus Classification Working Group. The web-based RotaC tool can be freely accessed at http://rotac.regatools.be.