Mon Jul 29 08:46:30 EST 1991

We've had considerable discusssion about this problem within the DOE
genome program. Procedures for making/rendering cDNA libraries equimolar
have been devised in labs of Sherman Weissman at Yale (recent PNAS paper)
and Agiris Efstratiadis at Yale.  Computer search for references.

However you may be better off working with a vactor providing for directional
cloning of the cDNA, but not normalizing.  The formalization process can tend
to eliminate members of multigene family which are individually interesting.

1. Make your library in a vector supporting directional
2. In the reverse transcription, use saturating quantities of polydT
as this will tend to shorten the length of cloned polydA 3'insert tails.
3. Prepare blots of your library.
4. Probe first with ribosomal DNA probe to avoid picking these dominant clones.
5. Initially sequence 3' ends as these will tend to be most divergence
within members of a single multigene family, that is, their 3' untranslated
6. When you find that you've re-sequenced a 3'end, use it as a probe on
your library blot to eliminate that specie of abundant clones.
7. AND so forth.

In this way you'll make your way through from more abundant to less abundant
clones, with only about a 2 X resequencing of 3' ends.

A second library produced through random priming on mRNA will probably be usefull
It can be used as a target (its blot that is) for probing with the 5'
ends of probes derived from the first library.
Thus 5' segments not reverse transcribed in the first library may be found

Do a good search on methodologies.  There are a lot of good tricks in the
Marvin Stodolsky with the DOE Human Genome group.

More information about the Methods mailing list