Builds amino-acid k-mer sets, calculates Jaccard similarity, and finds the highest contiguous identity between each supplied epitope and any equal-length window of the query. It does not predict immune binding, allergenicity, cross-reactivity or clinical risk and cannot replace curated databases or laboratory testing.

Standardizes each supplied value against its supplied mean and standard deviation, flips markers whose favorable direction is lower, and computes a weighted composite mapped to a bounded 0–100 index. The references and weights come entirely from the caller; this is not a validated longevity score, diagnosis, prognosis or medical advice.

Compares one guide with caller-supplied candidate protospacers, weights mismatches in the guide’s final ten positions twice, applies a simple NGG PAM penalty, and ranks a transparent similarity score. It does not search a genome, model bulges, chromatin or nuclease-specific biology, and must not be used as a clinical or laboratory safety decision.

Multiplies per-contact transmission probability, effective contacts per day and infectious duration to obtain a simple basic reproduction number, then applies a susceptible fraction for an effective number. It is a classroom homogeneous-mixing calculation, not an outbreak estimate, fitted epidemiological model or public-health forecast.

Divides each available substrate mass by its required mass per product mass, selects the limiting substrate, and applies a caller-supplied process efficiency. It does not model yeast metabolism, kinetics, oxygen transfer, toxicity, regulation or actual fermentation performance.