If we apply a basic algorithm to a reading frame to scan it and look for START and STOP codons (to assemble a possible protein), we get cases, when we have multiple START codons and one STOP codon:
['A', 'S', 'M', 'A', 'P', 'M', 'Q', 'P', 'I', 'T', 'P', 'S', 'A', 'T', '_', 'T']
We see that we have two START codons (M) and one STOP (_) codons and if we generate a possible protein from this reading frame, we will get two results:
First one contains the second one in it and also has a START codon in it.
Question (from a programmer): Is there ever a need to generate sequences like that? Is it useful for some kind f statistics, or do we discard the first one and only use sequences that have one START and one STOP codon?