In article <1994Jun21.173724.1 at ugsc2a>, otello at sc2a.unige.ch writes...
>>I would like to know if anybody out there knows of a method
>to detect putative tRNA sequences in a given nucleotide sequence.
>I have an 900 bp sequence in which somebody mapped a tRNA a while ago,
>and I would like to verify if it is really there, since I plan some mutagenesis
>in this fragment.
>>Thanks a lot in advance
There is an old Zucker program that will give putative tRNAs from a sequence.
However, one must be aware of some basic "rules of thumb" to intrepret the
output reasonably. For example:
1. acceptor stem has seven base pairs
2. Anticodon stem has 5 basepairs
3. anticodon loop as 7 nucleotides: two pyrimidines on the 5', two purines on
the 3'; three base anticodon. If yeast, fungi, watch out for intervening
4. T stem has 5 base pairs; 5 bases in the loop; the sequence GTTCPu should
appear at the begining of this loop (unless a eukaryotic initiation: GATCPu.
If position 54 is T, then position 60 should be T or C.
5. If bacterial, the CCA will be encoded. If eukaryotic, it won't be; expect
the gene to be bounded on the 3'-side past the descriminator base by a run of
These are guides; MOST tRNAs will fit; but not all.
The Zucker program uses an algorithm to find base pairing of 7 nucleotides to
another 7 nucleotides located an appropriate distance downstream. Then it
starts to apply other criteria (T stem/loop size, etc.); it only filters on
the specific bases if you pick that option. However,it is a big help to scan a
large sequence. I know the program runs on VAX; other platforms may be
available. In the meantime, you might be able to use what I have listed above
to eyeball it, especially if it's eukaryotic (find the oligo-T).
c/o Bhandary at wccf.mit.edu