discrepancies between cDNA and protein sequences

[Michael R. Thompson]

In article <3aqa0d$ad2 at news.univ-rennes1.fr>, lepennec at univ-rennes1.fr (Le Pennec Jean-Paul) says:
>
>I have analysed a cDNA.....

In my experience, cDNA synthesis, like PCR, is error prone. It is therefore
usually best to sequence at least two independently derived clones 
(preferably three, of course). I recently isolated several clones encoding 
a functional enzyme and several near identical clones which produced 
immunologically recognizable, but enzymatically inactive, protein. I 
sequenced two clones encoding enzymatically-non-functional protein and one 
encoding enzymatically-functional protein. There were seven base differences
in ca. 1.8 kb between the three sequences. Four of the differences were 
outside the reading frame, one of the base differences caused a silent 
mutation, and the other two (one in each of the 'non-functional' clones) 
caused severely non-conserved changes in amino acid sequence... presumably 
sufficient to disrupt protein structure. The simplest explanation for these 
differences is error in base incorporation during reverse transcription.