Frequent Asked Questions - FAQ

Does Macrel detect post-translationally modified peptides?

Currently, Macrel only detects mature peptides. Masking sequences such as leaders or signal peptides is not yet supported.

Does Macrel detect peptides of any length?

Macrel targets peptides in the range of 10-100 residues. The upper bound is enforced: predicted small genes longer than 100 amino acids are filtered out, and the classifier will generally not return a meaningful result for longer sequences. The lower bound is a practical recommendation rather than a hard filter — Macrel will not crash on very short sequences, but predictions for peptides shorter than ~10 residues are unlikely to be meaningful.

What happens to sequences with ambiguous or non-standard residues?

Empty sequences and sequences containing non-standard amino acid codes (such as X, B, Z, U, O, or an internal *) are skipped during feature computation. Macrel emits a warning for each skipped sequence and continues with the remaining input rather than aborting.

Can Macrel work well on isolate genomes or isolate genome reads?

Macrel is able to handle isolated genomes, as contigs, short reads (in this case it will trim and assemble before starting the gene prediction) or directly on predicted peptide sequences (obtained through some alternative pipeline).