Reproducibility of experiments

David Steffen steffen at mbcr.bcm.tmc.edu
Thu Apr 15 09:28:44 EST 1993


F. Lehmann, Ph.D. says:
 "I ask for a description of what a kit does when I review papers.

And Don Forsdyke (Discussion Leader) replies:
 "Splendid! I wish more reviewers would do the name. But the real
 "responsibiliy lies with journal editors. Phrases such as "according
 "to the manufacturer' s instructions" just should not be accepted.

I think these sentiments are admirable, but I think they are
inconsisent with currently accepted practice.  One major issue being
ignored here is the belief on the part of the Journals that they need
to limit the lenth of papers to keep page counts down.  (Reading a few
"instructions to authors" will confirm this).  In one sense, it is
intellectually satisfying to take the moral high ground and state
that every paper should contain a methods section detailed enough to
allow one to reproduce the reported experiments, but there are
economic (and ecological?) issues which arise from the vast increase in
paper length that would result from this.

Current practice is to describe only novel methods, with other methods
described by reference.  The problems with this are:

1) One can be lead down a long chain of references to references to
references to ...

2) It is too easy to say "Southern blots were done by the techniques
of Southern (J. Mol. Biol. 98: 503-517, 1975) when you know you have
made dozens of changes from the original protocol.

3) As in the case discussed above, the reference may be more or less
unavailable.  "According to manufacturers instructions" is one
example, but references to journals which are not widely available,
books, meetings proceedings, etc., in fact raise the same problems.

I propose the following general solution:

There ought to be a general repository of methods.  This ought to be
available both in hard copy and electronically.  Whenever possible,
discussions in the methods section of a paper should be either a
reference to one of the standard methods, or modifications thereto.
References to references should not be allowed (e.g. redescribe ALL
modifications from the standard protocols).  Kit manufacturers should
be strongly encouraged to deposit kit instructions and descriptions in
this archive.

It has not escaped my notice that this proposal has a lot in common
with the growing practice of referring to either Maniatis or Methods
in Molecular Biology (the Red Book).


-- 
David Steffen
Department of Cell Biology, Baylor College of Medicine, Houston TX 77030
Telephone = (713) 798-6655, FAX = (713) 790-0545
Internet = steffen at bcm.tmc.edu



More information about the Jrnlnote mailing list