4823 reasons to make your own libraries

[TJ Murphy]

Thought you might be interested in this warning I stumbled into on the
ATCC server.

By the way, reason number 4824 comes from the experience of a colleague
who purchased, from one of those Southern Californian mol biol stores we
get pounds of mail from every week, a human liver library that turned
out to be rat.  He attached a letter of apology for his wasted year from
the store to his grant renewal.

By the way, he got renewed.

We are letting all researchers who may have received clones from the
IMAGE/ Life Technologies mouse brain library or who
may be interested in ordering that there is significant evidence that
this library is actually from rat tissue. The following is an
excerpt of an email from John Wootton (NCBI):

"The initial discovery was made by my collaborators at NCI, Vadim
Gladyshev and Dolph Hatfield, who, incidentally deserve
credit for this and are also concerned that the community of users are
made aware of the problem. They obtained human and
mouse IMAGE clones matching a new gene and are carefully producing
finished sequence from these clones. One clone was
from the "mouse brain" library - it turned out to have 10 percent
mismatches from the other mouse clones. They searched
dbEST with their extended sequence from this clone and found it was
identical within error to ESTs from two rat libraries.

They informed me of this Tuesday July 8. I ran comprehensive searches of
the dbEST sequences attributed to this library
against the NCBI nr database using gapped blastn. In all cases examined,
where both rat and mouse orthologs exist in nr, the
ESTs from this library were identical within error to rat and
significantly (typically 7-12 percent) different from mouse. The
simplest explanation is that the entire library (4823 sequences in dbEST
as of 7/8/97) is from rat.

I did further searches to test if any other "mouse" libraries in dbEST
may be rat cDNAs, using gapped blastn searches of
dbEST with authentic rat sequences, including mitochondrial DNA. The
only libraries in dbEST that were identical within error
to these rat sequences were this "mouse brain" library and libraries
labelled rat. The Life-Tech mouse embryo libraries are
evidently authentic mouse.

The dbEST sequences from this "mouse brain" library contain a high
proportion of mitochondrial DNA matches. All of these are
almost identical to the authentic rat mitochondrial sequences in
GenBank. There are a few linked polymorphisms in some ESTs,
which suggest that that the library is a mixture of DNA from two
different strains of rat in approximate 60:40 proportion. The
authentic mouse mitochondrial sequences are approximately 15 percent
different from any of these rat sequences. This library
also contains more than half of all the rat nuclear gene matches in
dbEST: the other rat libraries are much smaller.

If you want to confirm these findings, here is a sample of GenBank rat
sequence accessions to use in blastn searches against
dbEST: M27315, J01435, V01574 (mitochondrial DNA from different
strains), M25889, J01435, M14400, J02744,
X05882, X60469, X13933, M19533, M24361, L15556 (or most other rat
housekeeping nuclear gene sequences)."