Skip to content

Glossary

Each term is defined briefly here and in full on the page it links to. Names in code are the fields of the JSON and TSV outputs or the options of the command line.

One column of the multiple sequence alignment: the residue, or gap, that each sequence has at that place. Columns are numbered from 1. See the input.

The window of k columns that starts at a given column; a position is numbered by its first column, so an alignment of L columns has L − k + 1 positions. See k-mer positions.

A k-mer is the k characters of one sequence inside a position’s window. k is set with -k (default 9 for proteins and 27 for nucleotides). See choose the k-mer length.

A k-mer whose characters are all residues. A gap, missing-data symbol, ambiguity code, stop or frameshift symbol anywhere in the window makes it invalid, and that sequence then does not count at that position. See valid k-mers and support.

A distinct valid k-mer at a position, with the number of sequences that carry it.

The number of sequences with a valid k-mer at a position (N, support).

The value T that each position’s support is compared with (-s, default 100, the paper’s value). See the support threshold.

A position with support above the support threshold. Only eligible positions receive the paper’s corrected entropy and enter the summary statistics.

The labels (low_support) of positions that are not eligible: NS, no support (N = 0); LS, low support (N below T); ELS, support equal to the threshold (N = T).

The Shannon entropy, in bits, of a position’s variant counts as observed (observed_entropy). It underestimates the entropy of the population, more so with fewer sequences. See observed entropy.

The entropy reported as entropy: by default the paper’s sample-size correction at eligible positions, and the observed entropy at the other positions with support. See the paper’s correction.

The method behind the reported entropy (--estimator): dima (the paper’s correction, the default), chao-wang-jost, miller-madow or raw. Each position records the method used (entropy_method). See alternative estimators.

Why a position’s entropy was computed with another method than the one requested (entropy_fallback), for example because its support is at or below the support threshold.

The estimated probability that one more sequence would carry a k-mer already seen at that position (sample_coverage). Low coverage means that much of the diversity there has probably not been sampled. See sample coverage.

The motif of each variant of a position: Index, the most common variant (count above 1); Major, the next most common; Minor, the other variants seen more than once; Unique, the variants seen once. See motif classes.

A variant’s count as a percentage of the position’s support.

The percentage of a position’s support carried by variants other than the Index.

The number of distinct non-Index variants as a percentage of the sequences that carry them. See incidences.

A region of consecutive positions whose Index variant reaches the threshold incidence (--hcs-min-incidence, default 100%). See historically conserved sequences.

A conserved region that includes positions with support at or below the support threshold, so that its conservation rests on few sequences. See flagging.

The exact 95% Clopper–Pearson lower bound of the Index proportion of a position or region (conservation_lower_bound_95): how conserved it is at least, given the number of sequences. See a lower bound on conservation.

The names of the |-separated fields of the FASTA headers (--header-format). See describe the headers.

The header fields per whose values each variant’s sequences are counted (--metadata-fields). See stratify by header metadata.

The value that replaces an empty header field (--header-fillna, default Unknown).

What happens to characters outside the alphabet (--validation): strict stops the run, permissive excludes the k-mers that contain them, and report also lists every character found. See validate input characters.

The name of the analysed protein or gene in the results (query_name): --query-name, or the input file name without its extensions. See the query name.

A condition that did not stop the analysis but that you should know about, reported once with a code, a count, examples and a hint. See Diagnostics.

The run object of the results: the version, every parameter and where it came from, the input’s checksum, what reading found, the diagnostics and the references, with nothing that varies between runs. See run.

An optional, separate JSON file (--run-report) that records how a run went: times, threads, memory, platform and exit status, without user or host names. See keep a run report.

The version of a JSON output’s format (schema_version). Its major number changes when the meaning of existing fields changes. See JSON outputs.