On 1/15/2011 12:28 PM, REX PALMER wrote:
Does anyone know of a statistical breakdown of successful protein
structure determinations in terms of the method used?
Rex Palmer
Birkbeck College
I think this was discussed back in April under "Proportion of MR in PDB":
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=CCP4BB;b5186393.1004
I did a little analysis then where I manually distilled the 860 unique
strings found under the "METHOD USED" entry of the PDB down into what I
think are distinct "methods":
37851 MR
16435 "NULL", "N/A", or no "REMARK 200" record at all
7802 MAD/SAD
993 MIR
669 OTHER
352 SIRAS
316 MIRAS
188 AB-INITIO
88 SIR
6 RIP
2 UNCONVENTIONAL
1 UNCONVENTIANAL
1 N?
1 FIBER-DIFFRACTION
64776 total
Note, that the second most popular "method" seems to be no method at
all, so I suppose that puts some large "error bars" on these numbers.
Although it is somewhat disturbing that such a large number of people
didn't seem to think their "method" was worth mentioning in their PDB
deposit, I don't this is because "methods" in general are considered
unimportant. It is because the "method" is considered obvious. There
are a lot of crystallographers out there who only know one "method". I
am often amused when I witness a meeting of two crystallographers who
are each shocked to learn that the other not only doesn't use their
favorite computer program, but has also never heard of it!
I am willing to bet that the earliest "no method" entries (particularly
the ones that lack a REMark 200 record) were probably MIR, since that
was the "obvious" method to solve a structure for some time. Modern
"NULL" entries seem to be mostly what I call "molecular replacement",
which includes just refining from a native, etc. not necessarily
running an MR search program. Or, at least, there are only ~26700
entries that explicitly state that the "method used to determine the
structure" was "molecular replacement", the rest I assigned to this
"bin" because the entry mentions a "starting model", etc.
In general, there is definitely confusion about nomenclature! For
example some structures that claim to be "Ab-initio" are actually
MAD/SAD structures. Although historically I think "Ab-initio" is
supposed to be synonymous with direct-methods based structure
determinations.
-James Holton
MAD Scientist
