How to interpret reports

Use this guide to interpret the most important report pages. Note that no single metric should be used alone to determine whether a viral detection is credible.

Interpreting the Classification table

The Classification table provides the main overview of viral taxa detected in a sample. Each row represents a viral taxon identified during the primary classification step and summarizes the results of subsequent validation and filtering steps.

The individual columns should not be interpreted in isolation. In particular, users should consider the primary classification together with BLASTn-based validation, read-assignment quality and genome coverage.

ColumnDescription
TaxID NCBI taxonomy identifier assigned during the primary classification.
Species name Viral species identified during the primary classification step.
BLAST hit Best matching taxonomic assignment obtained during BLASTn-based validation. This field can contain a more specific assignment than the species-level BLAST classification and is particularly useful when the BLAST species field does not agree with the primary classification.
BLAST species Species-level taxonomic assignment derived from the BLAST result. This value is compared with the Species name from the primary classification during the consolidation step.
Read number Number of reads assigned to the taxon during the primary classification step. This number can differ from the number of reads reported under Coverage → Statistics, because the latter is calculated after reads are mapped to the selected viral reference sequences.
Reclassification Result of the pipeline's amino-acid-based secondary taxonomic reclassification step.
Filtering Result of the coverage-based filtering step. A viral candidate passes when its read alignments satisfy the required coverage criteria, including a minimum number of non-adjacent covered regions, a minimum region length and a minimum number of covered bases. At least 3 non-adjacent regions, regions of at least 150 nt and at least 500 covered bases are required.
Consolidation Result of the BLASTn-based taxonomic comparison. The consolidation step passes when the species name obtained by BLASTn matches the species name reported during primary classification.
Coverage Percentage of the selected viral reference genome covered by mapped reads.
High quality read assignment Percentage of mapped reads with mapping quality ≥60. This metric indicates how confidently the reads can be assigned to the reported viral species rather than to closely related reference sequences.

High quality read assignment

The High quality read assignment value represents the proportion of reads mapping to the reported virus with a mapping quality of at least 60.

A value close to 100% indicates that almost all mapped reads could be assigned to the reported viral reference with high confidence. Lower values indicate that a substantial proportion of the reads also map to other, often closely related, reference sequences.

For example, a virus with a low high-quality read-assignment value may share most of its mapped reads with another related virus for which genome coverage is also reported. Such cases should therefore be interpreted together with the other detected taxa, their genome coverage and the corresponding BLASTn results.

This metric should not be interpreted as a probability that the reported virus is present. Rather, it describes the uniqueness and confidence of the read mappings to the selected reference sequence.

Why the read numbers can differ between Classification and Coverage → Statistics

The two report sections display read counts obtained at different stages of the analysis.

The Read number shown in the Classification table originates from the primary taxonomic classification.

By contrast, the Number of reads reported under Coverage → Statistics is calculated after reads have been mapped against the viral reference sequences selected as the most likely candidates during downstream analysis.

Consequently, the two values are not expected to be identical.

The Classification read count therefore reflects the evidence obtained during initial taxonomic assignment, whereas the Coverage read count describes how many reads subsequently mapped to the selected viral reference sequence used for calculating genome coverage and depth.

For interpretation of genome coverage, users should use the read number reported under Coverage → Statistics.

Filtering

The Filtering column reports the result of a coverage-based validation step designed to exclude viral assignments that are supported only by a small or highly localized part of a reference sequence.

A candidate must meet three criteria:

  • at least 3 non-adjacent covered regions,
  • a minimum covered-region length of 150 nt, and
  • at least 500 covered bases in total.

A PASSED result indicates that the read distribution fulfills these coverage criteria. A FAILED result indicates that the available read support is too limited or too localized to satisfy the selected thresholds.

Consolidation

Consolidation represents the result of the BLASTn-based taxonomic comparison.

The step is reported as PASSED when the species-level assignment obtained from BLASTn is exactly identical to the species identified during primary classification.

A FAILED consolidation does not necessarily mean that the original viral assignment is incorrect.

Taxonomic inconsistencies or differences in how records are represented in the NCBI taxonomy can occasionally result in different names being returned for biologically equivalent or closely related assignments. In addition, BLASTn-derived species-level taxonomy can in some cases correspond to the host species rather than directly to the viral taxon.

For example, the primary classification may identify Feline leukemia virus, whereas the species-level BLASTn assignment might be reported as Felis catus. In such a case, the exact species-name comparison causes Consolidation to fail even though the BLAST hit itself may correspond to the expected feline leukemia virus sequence.

Therefore, whenever Consolidation = FAILED, users should inspect both:

  • BLAST species, and
  • BLAST hit

before dismissing the viral assignment.

The BLAST hit can contain taxonomic information below the species level and may therefore provide a more informative interpretation than the species-level taxonomy alone in these edge cases.

Interpreting Coverage → Statistics

The Coverage → Statistics page describes how reads map across the selected viral reference genome. These metrics provide important supporting information for evaluating a viral detection.

MetricDescription
Number of reads Number of reads mapped to the selected viral reference during the downstream mapping step. This value may differ from the primary-classification read count.
Covered bases Number of nucleotide positions in the reference genome covered by at least one mapped read.
Percent covered Percentage of the reference genome covered by mapped reads.
Mean coverage Average sequencing depth across the complete reference genome.
Mean baseQ Mean base-quality score of the mapped bases.
Mean mapQ Mean mapping-quality score of the reads aligned to the reference genome.

Why genome coverage matters

Read count alone does not describe how sequence evidence is distributed across a viral genome.

For example, hundreds of reads may map to only a small region of a reference genome, whereas a smaller number of reads may be distributed across most of another viral genome. These two situations provide very different types of evidence.

The Percent covered metric therefore provides important context for interpreting the reported read count.

A high percentage of genome coverage indicates that sequencing reads support multiple regions across the viral reference. A low percentage indicates that the available evidence is restricted to a smaller part of the genome.

Low genome coverage does not by itself exclude the presence of the virus. Low-abundance viruses, short sequencing runs or uneven sequencing depth can all result in partial genome coverage. Such detections should, however, be interpreted more cautiously and together with the classification, BLAST and mapping-quality information.