Randy,

I believe that the first publication to report the use of coefficients 2Fobs-Fc 
antidates both the Main publication and the use of a downweighting of the Fc 
and is in Freer, et al., 1975 J. Biol. Chem. 250:46-54, which also gives a 
heuristic account of the value of such coefficients. 

----- Original Message -----
From: Randy Read <[email protected]>
Date: Friday, January 9, 2009 7:26 am
Subject: Re: [ccp4bb] Published derivation of mFo-DFc formula?
To: [email protected]

> Hi Ian,
> 
> Indeed, I didn't have time to reply during the festive season.
> 
> Just to give appropriate credit, the 2mFo-DFc coefficients 
> for  
> acentrics in SIGMAA were derived by analogy to arguments made by 
> Peter  
> Main, but taking account of the effect of errors in the 
> partial  
> model.  According to that argument, one doesn't expect 
> model bias in  
> the mFo coefficients for centrics, which is why they have 
> different  
> coefficients.
> 
> In the original version of SIGMAA, the difference map 
> coefficients  
> were mFo-Fc, not mFo-DFc.  At that time, I was thinking 
> that mFo gives  
> the best (lowest rms error, though biased) representation of the 
> true  
> density, while Fc represents the model so the difference between 
> them  
> should show most clearly how you need to change the model.  
> A year or  
> two later, I decided that it was more appropriate to smear out 
> the  
> model density according to its uncertainty, hence the addition 
> of the  
> D factor.  There were two more advantages to the D 
> factor.  First, if  
> the model is complete garbage, both m and D are zero so the 
> difference  
> map is flat.  (If you don't know anything about the true 
> phases, you  
> can't make a map showing you how to change it!)  Second, it 
> turns out  
> that if Fo is not on the same scale as Fc, the D term absorbs 
> the  
> necessary scale factor correction.  (Before I added the D 
> term,  
> Eleanor Dodson had written a comment in the source code of the 
> CCP4  
> version of SIGMAA, saying something along the lines "Randy seems 
> to be  
> assuming that the data are
> on absolute scale.  What on earth is he thinking?")
> 
> These arguments for the difference map coefficients are based 
> simply  
> on intuition about what makes sense to show how the model 
> should  
> change.  So I was very pleased, when we were working on 
> refinement  
> likelihood targets, to see that they can also be justified in 
> terms of  
> log-likelihood-gradient maps.  If you take the derivative 
> of the log- 
> likelihood functions with respect to Fc, you get
> 
> f(|s|)(mFo-DFc)
> 
> where f(|s|) is a resolution-dependent function given by
> 
> 2D/sigma-delta^2 for acentrics, and
>   D/sigma-delta^2 for centrics,
> 
> where sigma-delta^2 is the variance from the Rice function.
> 
> The D/sigma-delta^2 part would give a map that is the 
> convolution of a  
> map computed with the coefficients you suggest (i.e. 2mFo-2DFc 
> for  
> acentrics, mFo-DFc for centrics) and some shape function, which 
> might  
> sharpen the map if D/sigma-delta^2 increases with resolution or 
> smear  
> it out, if it decreases with resolution.
> 
> Anyway, at the least the factor of two for acentrics should 
> be  
> included in the various programs that compute difference 
> map  
> coefficients.  Someone should probably look at the effect 
> of the  
> convolution with the resolution-dependent part.  You will 
> get  
> different results if you consider the effect of coordinate error 
> to be  
> part of the model (e.g. D is taken up in the model partly 
> by  
> increasing B-factors -- if you take the derivative with respect 
> to  
> DFc, the factor D is missing from f(|s|)) or if you work in 
> terms of E- 
> values.  It's possible that the LLG map computed when the 
> likelihood  
> is expressed in terms of E-values would be optimal in terms of 
> being  
> sharpened as much as can be justified by the level of phase 
> error at  
> different resolutions.
> 
> Thanks for stimulating the discussion about this point!
> 
> Regards,
> 
> Randy Read
> 
> On 8 Jan 2009, at 11:43, Ian Tickle wrote:
> 
> >
> > All - I didn't get a single response to my posting last week
> > (https://www.jiscmail.ac.uk/cgi-
> bin/webadmin?A2=ind0812&L=CCP4BB&T=0&O=D> 
> &X=512817322E87355F7F&Y=i.tickle%40astex-therapeutics.com&P=266420)
> > concerning the formulae that are widely used for the 
> 'minimally- 
> > biased'
> > Fourier and difference Fourier coefficients.  It probably 
> didn't help
> > that I posted it in the middle of the festive season! - but still
> > somewhat surprising since I imagine everyone here is involved 
> with  
> > maps
> > at one time or another, and has an interest in getting the 
> density  
> > that
> > shows best what if any further modifications need to be made 
> to the
> > current model.  Anyway now that people have hopefully 
> returned to work
> > from the rigours of the CCP4 Study Weekend I thought I'd post 
> it again
> > and see if I can provoke some discussion this time.  I 
> won't post  
> > all my
> > calculations again, just a summary of my conclusions.
> >
> > First, I think I can now prove my conjecture that the 
> optimal  
> > difference
> > Fourier coefficient dF is given for both acentrics and 
> centrics by:
> >
> >     dF = Fm - DFc
> >
> > where Fm is the 'minimally-biased' Fourier coefficient derived 
> by Read
> > (AC 1986,A42,140):
> >
> >     Fm(acen) = 2mFo - DFc
> >     Fm(cen)  =  mFo
> >
> > I'm satisfied now that my alternative conjecture, that dF = Fm 
> - Fc,  
> > is
> > probably wrong.  Also I can see that there might be an 
> argument to put
> > DFc in the FC (FC_ALL) column in place of Fc as appears to 
> be  
> > currently
> > done by REFMAC, but not by SIGMAA (but I'd still like to see some
> > discussion of that).
> >
> > So here's a summary comparison of theory with what is my 
> understanding> is actually implemented in software, and with the 
> inconsistencies> highlighted (>...<):
> >
> >     Source       
> Coefficient   Acentrics        Centrics
> >     ======       
> ===========   =========        ========
> >
> >     THEORY(Read)     
> Fm        2mFo - 
> DFc       mFo
> >           
> ..  (me)       
> dF        2(mFo-
> DFc)       mFo - DFc
> >
> >     SIGMAA           Fm        2mFo - DFc       mFo
> >                          dF       > mFo - DFc <     mFo - DFc
> >                       Fc            Fc              Fc
> >
> >     REFMAC           Fm        2mFo - DFc    > 2mFo - DFc <
> >                          dF       > mFo - DFc <     mFo - DFc
> >                       Fc         > DFc <         > DFc <
> >
> > Even if you don't accept my suggestion for the acentric dF 
> coefficient> there are clearly some significant inconsistencies 
> between the
> > coefficients output by SIGMAA & REFMAC which it would be nice to
> > resolve!
> >
> > Cheers
> >
> > -- Ian
> >
> >
> > Disclaimer
> > This communication is confidential and may contain 
> privileged  
> > information intended solely for the named addressee(s). It may 
> not  
> > be used or disclosed except for the purpose for which it has 
> been  
> > sent. If you are not the intended recipient you must not 
> review,  
> > use, disclose, copy, distribute or take any action in reliance 
> upon  
> > it. If you have received this communication in error, please 
> notify  
> > Astex Therapeutics Ltd by emailing i.tic...@astex-
> therapeutics.com  
> > and destroy all copies of the message and any attached documents.
> > Astex Therapeutics Ltd monitors, controls and protects all 
> its  
> > messaging traffic in compliance with its corporate email 
> policy. The  
> > Company accepts no liability or responsibility for any 
> onward  
> > transmission or use of emails and attachments having left the 
> Astex  
> > Therapeutics domain.  Unless expressly stated, opinions 
> in this  
> > message are those of the individual sender and not of 
> Astex  
> > Therapeutics Ltd. The recipient should check this email and 
> any  
> > attachments for the presence of computer viruses. Astex 
> Therapeutics  
> > Ltd accepts no liability for damage caused by any virus 
> transmitted  
> > by this email. E-mail is susceptible to data corruption,  
> > interception, unauthorized amendment, and tampering, 
> Astex  
> > Therapeutics Ltd only send and receive e-mails on the basis 
> that the  
> > Company is not liable for any such alteration or any 
> consequences  
> > thereof.
> > Astex Therapeutics Ltd., Registered in England at 436 
> Cambridge  
> > Science Park, Cambridge CB4 0QA under number 3751674
> 
> ------
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical 
> Research      Tel: + 44 1223 336500
> Wellcome Trust/MRC 
> Building                   Fax: + 44 1223 336827
> Hills 
> Road                                    E-mail: [email protected]
> Cambridge CB2 0XY, 
> U.K.                       www- 
> structmed.cimr.cam.ac.uk
>

Reply via email to