Home Software Services About Contact usearch manual
accept options

See also
 
Search flowchart

  Termination options

Accept criteria determine whether an alignment is a hit, also called an accept. See also weak hits. Hits are written to the output files. The ‑maxhits N and ‑top_hits_only options specify that only the best hits are to be reported. Note that two or more hits may be tied for the best score or identity. Accepted hits are written to an output file sorted by decreasing alignment score (local alignments) or by decreasing identity (global alignments).

In clustering commands (cluster_fast and cluster_smallmem), accept options determine whether or not a sequence matches a cluster centroid and should be assigned to that cluster. A sequence can match only one centroid; this is usually the first accepted centroid, but this can be changed by increasing the -maxaccepts, in which case it will be the centroid with highest identity (see termination options).

Accept criteria do not have default values. If a given accept option is not specified, then the corresponding value is not computed or tested. So for example if -id is the only option given, then identity is the only value that is calculated from the alignment.

If more than one accept option is specified, they are combined with AND, so all of them must be satisfied.

Criteria that do not require an alignment, e.g. ‑idprefix and ‑minqt, are tested before an alignment is computed; these can give significant improvements in speed because a target can be rejected without the overhead of computing an alignment. Most of these are not supported by local search commands (ublast, usearch_local and search_local).

Option Real/
Integer
Global/
Local
Need aln? Description
evalue R L Y Maximum E-value. Required for most commands that use local alignments.
id R GL Y Minimum identity. Required for most commands that use global alignments.
‑query_cov R GL Y Fraction of the query sequence that is aligned, in the range 0.0 to 1.0.
‑target_cov R GL Y Fraction of the target sequence that is aligned, in the range 0.0 to 1.0.
‑idprefix I G N First N letters are identical.
‑idsuffix R G N Last N letters are identical.
‑minqt R G N Minimum value of query_seq_length / target_seq_length.
‑maxqt R G N Maximum value of query_seq_length / target_seq_length.
‑minsl R G N Minimum value of shorter_seq_length / longer_seq_length.
‑maxsl R G N Maximum value of shorter_seq_length / longer_seq_length.
‑leftjust   G Y No terminal gaps at start of alignment.
‑rightjust   G Y No terminal gaps at end of alignment.
‑self   GL N v.6.0.280 or later. Reject if labels are identical (i.e., reject self-hits).
‑selfid   G N v.6.0.280 or later. Reject if sequences are identical (i.e., don't want self-hits).
‑maxid R GL Y v6.0.284 and later. Reject if identity is greater. Example: to select hits that are 97% identical to two significant figures use ‑id 0.965 ‑maxid 0.975.
‑minsizeratio R GL N v7.0.1009 and later. Minimum query_size / target_size (see size annotations).
‑maxsizeratio R GL N v7.0.1009 and later. Maximum query_size / target_size (see size annotations).
‑maxdiffs I GL Y Maximum number of differences between the sequences, i.e. the maximum edit distance. A difference is defined to be an alignment column containing a gap or a substitution.
-maxsubs I GL Y v7.1 and later. Maximum number of alignment columns containing substitutions.
‑maxgaps I GL Y v7.1 and later. Maximum number of alignment columns containing gaps.
‑mincols I GL Y Minimum alignment length, i.e. minimum number of columns in the alignment.
‑maxqsize I GL N v7.0.1009 and later. Maximum query size annotation.
‑mintsize I GL N v7.0.1009 and later. Minimum target size annotation.
‑mid R GL Y v7.0 and later. Minimum match percent identity, defined as (number of columns containing identities) / (number of columns containing letter pars). Gapped columns are ignored. This is percent identity, not fractional identity like -id, so is in the range 0.0 to 100.0.