Primer3 scoring

Primer3 performs candidate selection and pair ranking. smfprimer maps its design parameters directly to Primer3 size, Tm, GC, and product-size settings.

Reading scores

pair_score is Primer3's pair penalty. Lower values rank ahead of higher values. The individual forward and reverse primer penalties are retained in the Python data model and in GenBank primer annotations.

The penalty is an objective function, not a probability or percentage. It reflects departures from preferred values and Primer3's pair constraints. Scores are most useful for comparing pairs from the same run with the same parameters.

Existing primers

evaluate_primer_pairs invokes Primer3's check_primers task. Supplied primers are retained even when they violate a configured threshold, and the corresponding Primer3 problems are returned as warnings.

Conversion-aware limitation

Primer3 scores the concrete expected converted allele rather than every member of a degenerate oligo mixture. Always inspect ambiguity counts. Highly degenerate primers can also exceed the Bowtie expansion cap used during specificity assessment.

--max-degeneracies (default: 3) discards candidate primers whose ambiguous Y/R base count exceeds the threshold before pairs are ranked, since Primer3 itself has no notion of ambiguity. The CLI accepts any non-negative integer; to disable the filter entirely, use the Python API with DesignParameters(max_degeneracies=None).