Re: [ccp4bb] Should Rmerge be reported?

[Winter, Graeme (DLSLtd,RAL,LSCI)]

Tim,

can I explain? To be honest no! Other than old habits and all. Should converge 
to the same value in the end, so you have a valid point.

All the best Graeme



From: Tim Gruene
Sent: Thursday, June 10, 2021 14:02
To: Winter, Graeme (DLSLtd,RAL,LSCI)
Cc: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Should Rmerge be reported?

Hi Graeme,

could you explain why Rmeas does not serve the same purpose as Rmerge?
I guess Manfred (and others) have no objection to reporting Rmeas just
instead of Rmerge.

@ Christy: If one of my manuscript were rejected solely because Rmerge
was not mentioned, I would make a phone call to the boss if the editor.
Afterall, Rmerge can be recovered from the unmerged data, which ideally
you did deposit at the PDB.

Best,
Tim


On Thu, 10 Jun 2021 12:42:10 + "Winter, Graeme (DLSLtd,RAL,LSCI)"
 wrote:

> Once again I find myself jumping to the defence of this rather poor
> statistic!
>
> Yes, Rmerge is a very poor estimator of "data quality" and has many
> well published flaws related to multiplicity, but the low resolution
> Rmerge, if combined with a multiplicity > (say) 5, is a good
> indicator of whether the data set is "good" or there is something odd
> going on.
>
> For example, if you claim a 1.6A structure with an inner shell Rmerge
> of 0.11, 5-fold multiplicity and an overall I/sig(I) of 68 I would
> "smell a rat"
>
> To me it does have a value, as an unbiased estimator of your true
> unmerged I/sigma as it does not depend on any manipulation you have
> done to your sigmas. It is not a good estimator of where the
> resolution should be cut or any other decisions.
>
> The above situation could be an indicator that there was radiation
> damage, for example
>
> There are better ways of measuring damage - Rd, Rcp, ... but these
> are not commonplace graphs as I understand it. This little number in
> the middle of the table does give you that hint.
>
> So while I would say rejecting a paper because it was not included
> was very heavy handed, I would not like to see it erased from all
> papers either.
>
> All the best Graeme
> ____
> From: CCP4 bulletin board  on behalf of
> Manfred S. Weiss  Sent: 10 June
> 2021 13:30 To: CCP4BB@JISCMAIL.AC.UK 
> Subject: Re: [ccp4bb] Should Rmerge be reported?
>
> Dear Cristy,
>
> this is really hilarious. And it just shows how attached
> some ppl are to outdated numbers. Against better
> knowledge.
>
> It has been shown many times that Rmerge is flawed
> at various levels.
>
> The only reason I can see to report it is to be backwards
> compatible. But of course, this is a really weak reason.
>
> I would love to see it disappear.
>
> All the best
> Manfred
>
> Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
> Dear Colleagues
> Hope to find you all well and healthy.
>
> I have a question regarding Rmerge. In recent years, we have
> published our crystallographic structures in highly respected
> journals using CC1/2, I/sigma(I), completeness and multiplicity as
> quality parameters for our diffraction data.
>
> Recently this year, We submitted a paper using the same strategy, but
> one of the reviewers asked us to provide the Rmerge, arguing that
> providing this data was compulsory and it was important to estimate
> radiation damage.
>
> We replied to the editor arguing that Rmerge should not be used as a
> quality parameter, as suggested by more recent literature, such as
> the article published by Karplus and Diederichs
> (10.1016/j.sbi.2015.07.003). We also argued that there are modern and
> efficient methods to estimate radiation damage (
> doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>;
> doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>).
> It is my opinion that an experienced crystallographer can even
> monitor radiation damage over the course of data processing.
>
> And our paper was rejected  due to the fact I did not
> provide Rmerge which I certainly could have done If I found necessary.
>
> Journals like Nature ((https://www.nature.com › documents ›
> nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
> and even IUCr Journals
> (https://journals.iucr.org/f/services/structuralcommunications/)
> still list Rmerge as a data to be reported. I always took this as a
> suggestion since there are people still using Rmerge for data cutoff,
> but I never took this as

Re: [ccp4bb] Should Rmerge be reported?

[Jon Cooper]

Correct me if I'm wrong, but the first papers to stop reporting R-anything were 
XFEL ones so I assumed it was a particularly bad estimator of data quality for 
those expt's ;-

With a more normal expt, again correct this if it's wrong, the problems with 
R-merge only throw it off by a few percent?

Cheers, Jon.C.

Sent from ProtonMail mobile

 Original Message 
On 10 Jun 2021, 14:07, Frank von Delft wrote:

> Or just boycott the journal...??
>
> On 10/06/2021 14:02, Tim Gruene wrote:
>> Hi Graeme,
>>
>> could you explain why Rmeas does not serve the same purpose as Rmerge?
>> I guess Manfred (and others) have no objection to reporting Rmeas just
>> instead of Rmerge.
>>
>> @ Christy: If one of my manuscript were rejected solely because Rmerge
>> was not mentioned, I would make a phone call to the boss if the editor.
>> Afterall, Rmerge can be recovered from the unmerged data, which ideally
>> you did deposit at the PDB.
>>
>> Best,
>> Tim
>>
>>
>> On Thu, 10 Jun 2021 12:42:10 + "Winter, Graeme (DLSLtd,RAL,LSCI)"
>>  wrote:
>>
>>> Once again I find myself jumping to the defence of this rather poor
>>> statistic!
>>>
>>> Yes, Rmerge is a very poor estimator of "data quality" and has many
>>> well published flaws related to multiplicity, but the low resolution
>>> Rmerge, if combined with a multiplicity > (say) 5, is a good
>>> indicator of whether the data set is "good" or there is something odd
>>> going on.
>>>
>>> For example, if you claim a 1.6A structure with an inner shell Rmerge
>>> of 0.11, 5-fold multiplicity and an overall I/sig(I) of 68 I would
>>> "smell a rat"
>>>
>>> To me it does have a value, as an unbiased estimator of your true
>>> unmerged I/sigma as it does not depend on any manipulation you have
>>> done to your sigmas. It is not a good estimator of where the
>>> resolution should be cut or any other decisions.
>>>
>>> The above situation could be an indicator that there was radiation
>>> damage, for example
>>>
>>> There are better ways of measuring damage - Rd, Rcp, ... but these
>>> are not commonplace graphs as I understand it. This little number in
>>> the middle of the table does give you that hint.
>>>
>>> So while I would say rejecting a paper because it was not included
>>> was very heavy handed, I would not like to see it erased from all
>>> papers either.
>>>
>>> All the best Graeme
>>> 
>>> From: CCP4 bulletin board  on behalf of
>>> Manfred S. Weiss  Sent: 10 June
>>> 2021 13:30 To: CCP4BB@JISCMAIL.AC.UK 
>>> Subject: Re: [ccp4bb] Should Rmerge be reported?
>>>
>>> Dear Cristy,
>>>
>>> this is really hilarious. And it just shows how attached
>>> some ppl are to outdated numbers. Against better
>>> knowledge.
>>>
>>> It has been shown many times that Rmerge is flawed
>>> at various levels.
>>>
>>> The only reason I can see to report it is to be backwards
>>> compatible. But of course, this is a really weak reason.
>>>
>>> I would love to see it disappear.
>>>
>>> All the best
>>> Manfred
>>>
>>> Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
>>> Dear Colleagues
>>> Hope to find you all well and healthy.
>>>
>>> I have a question regarding Rmerge. In recent years, we have
>>> published our crystallographic structures in highly respected
>>> journals using CC1/2, I/sigma(I), completeness and multiplicity as
>>> quality parameters for our diffraction data.
>>>
>>> Recently this year, We submitted a paper using the same strategy, but
>>> one of the reviewers asked us to provide the Rmerge, arguing that
>>> providing this data was compulsory and it was important to estimate
>>> radiation damage.
>>>
>>> We replied to the editor arguing that Rmerge should not be used as a
>>> quality parameter, as suggested by more recent literature, such as
>>> the article published by Karplus and Diederichs
>>> (10.1016/j.sbi.2015.07.003). We also argued that there are modern and
>>> efficient methods to estimate radiation damage (
>>> doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>;
>>> doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>).
>&g

Re: [ccp4bb] Should Rmerge be reported?

[Frank von Delft]

Or just boycott the journal...??


On 10/06/2021 14:02, Tim Gruene wrote:

Hi Graeme,

could you explain why Rmeas does not serve the same purpose as Rmerge?
I guess Manfred (and others) have no objection to reporting Rmeas just
instead of Rmerge.

@ Christy: If one of my manuscript were rejected solely because Rmerge
was not mentioned, I would make a phone call to the boss if the editor.
Afterall, Rmerge can be recovered from the unmerged data, which ideally
you did deposit at the PDB.

Best,
Tim


On Thu, 10 Jun 2021 12:42:10 + "Winter, Graeme (DLSLtd,RAL,LSCI)"
 wrote:


Once again I find myself jumping to the defence of this rather poor
statistic!

Yes, Rmerge is a very poor estimator of "data quality" and has many
well published flaws related to multiplicity, but the low resolution
Rmerge, if combined with a multiplicity > (say) 5, is a good
indicator of whether the data set is "good" or there is something odd
going on.

For example, if you claim a 1.6A structure with an inner shell Rmerge
of 0.11, 5-fold multiplicity and an overall I/sig(I) of 68 I would
"smell a rat"

To me it does have a value, as an unbiased estimator of your true
unmerged I/sigma as it does not depend on any manipulation you have
done to your sigmas. It is not a good estimator of where the
resolution should be cut or any other decisions.

The above situation could be an indicator that there was radiation
damage, for example

There are better ways of measuring damage - Rd, Rcp, ... but these
are not commonplace graphs as I understand it. This little number in
the middle of the table does give you that hint.

So while I would say rejecting a paper because it was not included
was very heavy handed, I would not like to see it erased from all
papers either.

All the best Graeme

From: CCP4 bulletin board  on behalf of
Manfred S. Weiss  Sent: 10 June
2021 13:30 To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] Should Rmerge be reported?

Dear Cristy,

this is really hilarious. And it just shows how attached
some ppl are to outdated numbers. Against better
knowledge.

It has been shown many times that Rmerge is flawed
at various levels.

The only reason I can see to report it is to be backwards
compatible. But of course, this is a really weak reason.

I would love to see it disappear.

All the best
Manfred

Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have
published our crystallographic structures in highly respected
journals using CC1/2, I/sigma(I), completeness and multiplicity as
quality parameters for our diffraction data.

Recently this year, We submitted a paper using the same strategy, but
one of the reviewers asked us to provide the Rmerge, arguing that
providing this data was compulsory and it was important to estimate
radiation damage.

We replied to the editor arguing that Rmerge should not be used as a
quality parameter, as suggested by more recent literature, such as
the article published by Karplus and Diederichs
(10.1016/j.sbi.2015.07.003). We also argued that there are modern and
efficient methods to estimate radiation damage (
doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>;
doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>).
It is my opinion that an experienced crystallographer can even
monitor radiation damage over the course of data processing.

And our paper was rejected  due to the fact I did not
provide Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents ›
nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
and even IUCr Journals
(https://journals.iucr.org/f/services/structuralcommunications/)
still list Rmerge as a data to be reported. I always took this as a
suggestion since there are people still using Rmerge for data cutoff,
but I never took this as if Rmerge was a compulsory data to be
reported.

I would like to hear the opinion of this community. Should we
compulsorily report Rmerge?  If so, Why?

Cheers,

Cristy
--
Cristina Nonato
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo





To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1


--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin
Germany



Helmholtz-Zentrum Berlin für Materialien 

Re: [ccp4bb] Should Rmerge be reported?

[Tim Gruene]

Hi Graeme,

could you explain why Rmeas does not serve the same purpose as Rmerge?
I guess Manfred (and others) have no objection to reporting Rmeas just
instead of Rmerge.

@ Christy: If one of my manuscript were rejected solely because Rmerge
was not mentioned, I would make a phone call to the boss if the editor.
Afterall, Rmerge can be recovered from the unmerged data, which ideally
you did deposit at the PDB.

Best,
Tim


On Thu, 10 Jun 2021 12:42:10 + "Winter, Graeme (DLSLtd,RAL,LSCI)"
 wrote:

> Once again I find myself jumping to the defence of this rather poor
> statistic!
> 
> Yes, Rmerge is a very poor estimator of "data quality" and has many
> well published flaws related to multiplicity, but the low resolution
> Rmerge, if combined with a multiplicity > (say) 5, is a good
> indicator of whether the data set is "good" or there is something odd
> going on.
> 
> For example, if you claim a 1.6A structure with an inner shell Rmerge
> of 0.11, 5-fold multiplicity and an overall I/sig(I) of 68 I would
> "smell a rat"
> 
> To me it does have a value, as an unbiased estimator of your true
> unmerged I/sigma as it does not depend on any manipulation you have
> done to your sigmas. It is not a good estimator of where the
> resolution should be cut or any other decisions.
> 
> The above situation could be an indicator that there was radiation
> damage, for example
> 
> There are better ways of measuring damage - Rd, Rcp, ... but these
> are not commonplace graphs as I understand it. This little number in
> the middle of the table does give you that hint.
> 
> So while I would say rejecting a paper because it was not included
> was very heavy handed, I would not like to see it erased from all
> papers either.
> 
> All the best Graeme
> ____
> From: CCP4 bulletin board  on behalf of
> Manfred S. Weiss  Sent: 10 June
> 2021 13:30 To: CCP4BB@JISCMAIL.AC.UK 
> Subject: Re: [ccp4bb] Should Rmerge be reported?
> 
> Dear Cristy,
> 
> this is really hilarious. And it just shows how attached
> some ppl are to outdated numbers. Against better
> knowledge.
> 
> It has been shown many times that Rmerge is flawed
> at various levels.
> 
> The only reason I can see to report it is to be backwards
> compatible. But of course, this is a really weak reason.
> 
> I would love to see it disappear.
> 
> All the best
> Manfred
> 
> Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
> Dear Colleagues
> Hope to find you all well and healthy.
> 
> I have a question regarding Rmerge. In recent years, we have
> published our crystallographic structures in highly respected
> journals using CC1/2, I/sigma(I), completeness and multiplicity as
> quality parameters for our diffraction data.
> 
> Recently this year, We submitted a paper using the same strategy, but
> one of the reviewers asked us to provide the Rmerge, arguing that
> providing this data was compulsory and it was important to estimate
> radiation damage.
> 
> We replied to the editor arguing that Rmerge should not be used as a
> quality parameter, as suggested by more recent literature, such as
> the article published by Karplus and Diederichs
> (10.1016/j.sbi.2015.07.003). We also argued that there are modern and
> efficient methods to estimate radiation damage (
> doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>;
> doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>).
> It is my opinion that an experienced crystallographer can even
> monitor radiation damage over the course of data processing.
> 
> And our paper was rejected  due to the fact I did not
> provide Rmerge which I certainly could have done If I found necessary.
> 
> Journals like Nature ((https://www.nature.com › documents ›
> nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
> and even IUCr Journals
> (https://journals.iucr.org/f/services/structuralcommunications/)
> still list Rmerge as a data to be reported. I always took this as a
> suggestion since there are people still using Rmerge for data cutoff,
> but I never took this as if Rmerge was a compulsory data to be
> reported.
> 
> I would like to hear the opinion of this community. Should we
> compulsorily report Rmerge?  If so, Why?
> 
> Cheers,
> 
> Cristy
> --
> Cristina Nonato
> Associate Professor
> Laboratório de Cristalografia de Proteínas
> Faculdade de

Re: [ccp4bb] Should Rmerge be reported?

[Winter, Graeme (DLSLtd,RAL,LSCI)]

Dear Manfred,

I completely agree. However this is not routinely reported.

The discussion is of removing one indicator of this signal, rather than 
replacing it with a superior measure - the latter is something I would more 
actively support however would involve adding new stuff which is an uphill 
battle.

All the best Graeme

From: Manfred S. Weiss 
Sent: 10 June 2021 13:43
To: Winter, Graeme (DLSLtd,RAL,LSCI) ; 
CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] Should Rmerge be reported?

Dear Graeme,

a better number to look at is Isa. Instead of low resolution Rmerge.
Seriously. We monitor that to assess beamline performance.

Cheers, Manfred

Am 10.06.2021 um 14:42 schrieb Winter, Graeme (DLSLtd,RAL,LSCI):
Once again I find myself jumping to the defence of this rather poor statistic!

Yes, Rmerge is a very poor estimator of "data quality" and has many well 
published flaws related to multiplicity, but the low resolution Rmerge, if 
combined with a multiplicity > (say) 5, is a good indicator of whether the data 
set is "good" or there is something odd going on.

For example, if you claim a 1.6A structure with an inner shell Rmerge of 0.11, 
5-fold multiplicity and an overall I/sig(I) of 68 I would "smell a rat"

To me it does have a value, as an unbiased estimator of your true unmerged 
I/sigma as it does not depend on any manipulation you have done to your sigmas. 
It is not a good estimator of where the resolution should be cut or any other 
decisions.

The above situation could be an indicator that there was radiation damage, for 
example

There are better ways of measuring damage - Rd, Rcp, ... but these are not 
commonplace graphs as I understand it. This little number in the middle of the 
table does give you that hint.

So while I would say rejecting a paper because it was not included was very 
heavy handed, I would not like to see it erased from all papers either.

All the best Graeme

From: CCP4 bulletin board <mailto:CCP4BB@JISCMAIL.AC.UK> 
on behalf of Manfred S. Weiss 
<mailto:manfred.we...@helmholtz-berlin.de>
Sent: 10 June 2021 13:30
To: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> 
<mailto:CCP4BB@JISCMAIL.AC.UK>
Subject: Re: [ccp4bb] Should Rmerge be reported?

Dear Cristy,

this is really hilarious. And it just shows how attached
some ppl are to outdated numbers. Against better
knowledge.

It has been shown many times that Rmerge is flawed
at various levels.

The only reason I can see to report it is to be backwards
compatible. But of course, this is a really weak reason.

I would love to see it disappear.

All the best
Manfred

Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have published our 
crystallographic structures in highly respected journals using CC1/2, 
I/sigma(I), completeness and multiplicity as quality parameters for our 
diffraction data.

Recently this year, We submitted a paper using the same strategy, but one of 
the reviewers asked us to provide the Rmerge, arguing that providing this data 
was compulsory and it was important to estimate radiation damage.

We replied to the editor arguing that Rmerge should not be used as a quality 
parameter, as suggested by more recent literature, such as the article 
published by Karplus and Diederichs (10.1016/j.sbi.2015.07.003). We also argued 
that there are modern and efficient methods to estimate radiation damage ( 
doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>; 
doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>). 
It is my opinion that an experienced crystallographer can even  monitor 
radiation damage over the course of data processing.

And our paper was rejected  due to the fact I did not provide 
Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents › 
nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
 and even IUCr Journals 
(https://journals.iucr.org/f/services/structuralcommunications/) still list 
Rmerge as a data to be reported. I always took this as a suggestion since there 
are people still using Rmerge for data cutoff, but I never took this as if 
Rmerge was a compulsory data to be reported.

I would like to hear the opinion of this community. Should we compulsorily 
report Rmerge?  If so, Why?

Cheers,

Cristy
--
Cristina Nonato
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo





To

Re: [ccp4bb] Should Rmerge be reported?

[Manfred S. Weiss]

Dear Graeme,

a better number to look at is Isa. Instead of low resolution Rmerge.
Seriously. We monitor that to assess beamline performance.

Cheers, Manfred

Am 10.06.2021 um 14:42 schrieb Winter, Graeme (DLSLtd,RAL,LSCI):
Once again I find myself jumping to the defence of this rather poor statistic!

Yes, Rmerge is a very poor estimator of "data quality" and has many well published flaws 
related to multiplicity, but the low resolution Rmerge, if combined with a multiplicity > (say) 5, 
is a good indicator of whether the data set is "good" or there is something odd going on.

For example, if you claim a 1.6A structure with an inner shell Rmerge of 0.11, 5-fold 
multiplicity and an overall I/sig(I) of 68 I would "smell a rat"

To me it does have a value, as an unbiased estimator of your true unmerged 
I/sigma as it does not depend on any manipulation you have done to your sigmas. 
It is not a good estimator of where the resolution should be cut or any other 
decisions.

The above situation could be an indicator that there was radiation damage, for 
example

There are better ways of measuring damage - Rd, Rcp, ... but these are not 
commonplace graphs as I understand it. This little number in the middle of the 
table does give you that hint.

So while I would say rejecting a paper because it was not included was very 
heavy handed, I would not like to see it erased from all papers either.

All the best Graeme

From: CCP4 bulletin board <mailto:CCP4BB@JISCMAIL.AC.UK> on behalf 
of Manfred S. Weiss 
<mailto:manfred.we...@helmholtz-berlin.de>
Sent: 10 June 2021 13:30
To: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> 
<mailto:CCP4BB@JISCMAIL.AC.UK>
Subject: Re: [ccp4bb] Should Rmerge be reported?

Dear Cristy,

this is really hilarious. And it just shows how attached
some ppl are to outdated numbers. Against better
knowledge.

It has been shown many times that Rmerge is flawed
at various levels.

The only reason I can see to report it is to be backwards
compatible. But of course, this is a really weak reason.

I would love to see it disappear.

All the best
Manfred

Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have published our 
crystallographic structures in highly respected journals using CC1/2, 
I/sigma(I), completeness and multiplicity as quality parameters for our 
diffraction data.

Recently this year, We submitted a paper using the same strategy, but one of 
the reviewers asked us to provide the Rmerge, arguing that providing this data 
was compulsory and it was important to estimate radiation damage.

We replied to the editor arguing that Rmerge should not be used as a quality parameter, as 
suggested by more recent literature, such as the article published by Karplus and 
Diederichs (10.1016/j.sbi.2015.07.003). We also argued that there are modern and efficient 
methods to estimate radiation damage ( 
doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>; 
doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>). It is 
my opinion that an experienced crystallographer can even  monitor radiation damage over the 
course of data processing.

And our paper was rejected  due to the fact I did not provide 
Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents › 
nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
 and even IUCr Journals (https://journals.iucr.org/f/services/structuralcommunications/) still list Rmerge as a data 
to be reported. I always took this as a suggestion since there are people still using Rmerge for data cutoff, but I 
never took this as if Rmerge was a compulsory data to be reported.

I would like to hear the opinion of this community. Should we compulsorily 
report Rmerge?  If so, Why?

Cheers,

Cristy
--
Cristina Nonato
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo





To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1


--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin
Germany



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseh

Re: [ccp4bb] Should Rmerge be reported?

[Winter, Graeme (DLSLtd,RAL,LSCI)]

Once again I find myself jumping to the defence of this rather poor statistic!

Yes, Rmerge is a very poor estimator of "data quality" and has many well 
published flaws related to multiplicity, but the low resolution Rmerge, if 
combined with a multiplicity > (say) 5, is a good indicator of whether the data 
set is "good" or there is something odd going on.

For example, if you claim a 1.6A structure with an inner shell Rmerge of 0.11, 
5-fold multiplicity and an overall I/sig(I) of 68 I would "smell a rat"

To me it does have a value, as an unbiased estimator of your true unmerged 
I/sigma as it does not depend on any manipulation you have done to your sigmas. 
It is not a good estimator of where the resolution should be cut or any other 
decisions.

The above situation could be an indicator that there was radiation damage, for 
example

There are better ways of measuring damage - Rd, Rcp, ... but these are not 
commonplace graphs as I understand it. This little number in the middle of the 
table does give you that hint.

So while I would say rejecting a paper because it was not included was very 
heavy handed, I would not like to see it erased from all papers either.

All the best Graeme

From: CCP4 bulletin board  on behalf of Manfred S. Weiss 

Sent: 10 June 2021 13:30
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] Should Rmerge be reported?

Dear Cristy,

this is really hilarious. And it just shows how attached
some ppl are to outdated numbers. Against better
knowledge.

It has been shown many times that Rmerge is flawed
at various levels.

The only reason I can see to report it is to be backwards
compatible. But of course, this is a really weak reason.

I would love to see it disappear.

All the best
Manfred

Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have published our 
crystallographic structures in highly respected journals using CC1/2, 
I/sigma(I), completeness and multiplicity as quality parameters for our 
diffraction data.

Recently this year, We submitted a paper using the same strategy, but one of 
the reviewers asked us to provide the Rmerge, arguing that providing this data 
was compulsory and it was important to estimate radiation damage.

We replied to the editor arguing that Rmerge should not be used as a quality 
parameter, as suggested by more recent literature, such as the article 
published by Karplus and Diederichs (10.1016/j.sbi.2015.07.003). We also argued 
that there are modern and efficient methods to estimate radiation damage ( 
doi.org/10.1107/S1600576718005241<http://doi.org/10.1107/S1600576718005241>; 
doi.org/10.1107/S0907444909040177<http://doi.org/10.1107/S0907444909040177>). 
It is my opinion that an experienced crystallographer can even  monitor 
radiation damage over the course of data processing.

And our paper was rejected  due to the fact I did not provide 
Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents › 
nr-tables-xray<https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiigtb19uHwAhWS3YUKHSB1AdkQFjABegQIAxAD&url=https%3A%2F%2Fwww.nature.com%2Fdocuments%2Fnr-tables-xray.doc&usg=AOvVaw1RbfYvNeiEM07FBEOohMig>)
 and even IUCr Journals 
(https://journals.iucr.org/f/services/structuralcommunications/) still list 
Rmerge as a data to be reported. I always took this as a suggestion since there 
are people still using Rmerge for data cutoff, but I never took this as if 
Rmerge was a compulsory data to be reported.

I would like to hear the opinion of this community. Should we compulsorily 
report Rmerge?  If so, Why?

Cheers,

Cristy
--
Cristina Nonato
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo





To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1


--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin
Germany



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseher
Geschäftsführung: Prof. Dr. Bernd Rech (Sprecher), Prof. Dr. Jan Lüning, Thomas 
Frederking

Sitz Berlin, AG Charlottenburg, 89 HRB 5583

Postadresse:
Hahn-Meitner-Platz 1
14109 Berlin
Deutschland



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1

-- 
This e-mail and a

Re: [ccp4bb] Should Rmerge be reported?

[Manfred S. Weiss]

Dear Cristy,

this is really hilarious. And it just shows how attached
some ppl are to outdated numbers. Against better
knowledge.

It has been shown many times that Rmerge is flawed
at various levels.

The only reason I can see to report it is to be backwards
compatible. But of course, this is a really weak reason.

I would love to see it disappear.

All the best
Manfred

Am 10.06.2021 um 14:25 schrieb Maria Cristina Nonato:
Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have published our 
crystallographic structures in highly respected journals using CC1/2, 
I/sigma(I), completeness and multiplicity as quality parameters for our 
diffraction data.

Recently this year, We submitted a paper using the same strategy, but one of 
the reviewers asked us to provide the Rmerge, arguing that providing this data 
was compulsory and it was important to estimate radiation damage.

We replied to the editor arguing that Rmerge should not be used as a quality parameter, as 
suggested by more recent literature, such as the article published by Karplus and 
Diederichs (10.1016/j.sbi.2015.07.003). We also argued that there are modern and efficient 
methods to estimate radiation damage ( 
doi.org/10.1107/S1600576718005241; 
doi.org/10.1107/S0907444909040177). It is 
my opinion that an experienced crystallographer can even  monitor radiation damage over the 
course of data processing.

And our paper was rejected  due to the fact I did not provide 
Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents › 
nr-tables-xray)
 and even IUCr Journals (https://journals.iucr.org/f/services/structuralcommunications/) still list Rmerge as a data 
to be reported. I always took this as a suggestion since there are people still using Rmerge for data cutoff, but I 
never took this as if Rmerge was a compulsory data to be reported.

I would like to hear the opinion of this community. Should we compulsorily 
report Rmerge?  If so, Why?

Cheers,

Cristy
--
Cristina Nonato
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo





To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1


--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin
Germany



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseher
Geschäftsführung: Prof. Dr. Bernd Rech (Sprecher), Prof. Dr. Jan Lüning, Thomas 
Frederking

Sitz Berlin, AG Charlottenburg, 89 HRB 5583

Postadresse:
Hahn-Meitner-Platz 1
14109 Berlin
Deutschland



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Should Rmerge be reported?

[Maria Cristina Nonato]

Dear Colleagues
Hope to find you all well and healthy.

I have a question regarding Rmerge. In recent years, we have published our
crystallographic structures in highly respected journals using CC1/2,
I/sigma(I), completeness and multiplicity as quality parameters for our
diffraction data.

Recently this year, We submitted a paper using the same strategy, but one
of the reviewers asked us to provide the Rmerge, arguing that providing
this data was compulsory and it was important to estimate radiation damage.

We replied to the editor arguing that Rmerge should not be used as a
quality parameter, as suggested by more recent literature, such as the
article published by Karplus and Diederichs (10.1016/j.sbi.2015.07.003). We
also argued that there are modern and efficient methods to estimate
radiation damage ( doi.org/10.1107/S1600576718005241;
doi.org/10.1107/S0907444909040177). It is my opinion that an experienced
crystallographer can even  monitor radiation damage over the course of data
processing.

And our paper was rejected  due to the fact I did not
provide Rmerge which I certainly could have done If I found necessary.

Journals like Nature ((https://www.nature.com › documents › nr-tables-xray

) and even IUCr Journals (
https://journals.iucr.org/f/services/structuralcommunications/) still list
Rmerge as a data to be reported. I always took this as a suggestion since
there are people still using Rmerge for data cutoff, but I never took this
as if Rmerge was a compulsory data to be reported.

I would like to hear the opinion of this community. Should we compulsorily
report Rmerge?  If so, Why?

Cheers,

Cristy
-- 

*Cristina Nonato*
Associate Professor
Laboratório de Cristalografia de Proteínas
Faculdade de Ciências Farmacêuticas de Ribeirão Preto
University of São Paulo



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/