Recent work on gene concepts has been influenced by recognition of the extent to which RNA transcripts from a given DNA sequence yield different products in different cellular environments. These transcripts are altered in many ways and yield many products based, somehow, on the sequence of nucleotides in the DNA. I focus on alternative splicing of RNA transcripts (which often yields distinct proteins from the same raw transcript) and on 'gene sharing', in which a single gene produces distinct proteins with the exact same amino acid sequence. These are instances of molecular pleiotropy, in which distinct molecules are derived from a single putative gene. In such cases the cellular and external environments play major roles in determining which protein is produced. Where there is molecular pleiotropy, alternative gene concepts are naturally deployed; molecular epigenesis (revision of sequence-based information by altering molecular conformations or by action of non-informational molecules) plays a major role in orderly development. These results show that gene concepts in molecular biology do, and should, have both structural and functional components. They also show the need for a plurality of gene concepts and reveal fundamental difficulties in stabilizing gene concepts solely by reference to nucleotide sequence.