Map quality - phenixbb - phenix-online.org

newer
on B-factor refine of the...

Map quality

older
Several postdoc opportunities,...

mohamed noor

15 May 2015 15 May '15

4:10 p.m.

Dear all Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells. Thanks.

Attachments:

attachment.html (text/html — 308 bytes)

Reply

Sign in to reply online Use email software

Show replies by date

Pavel Afonine

15 May 15 May

4:37 p.m.

Hi Mohamed, this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues. Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map In a very nutshell, - "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC. All the best, Pavel On 5/15/15 9:10 AM, mohamed noor wrote:

Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks.

Reply

Sign in to reply online Use email software

Alexandre OURJOUMTSEV

6:43 p.m.

Dear Mohamed, I second Pavel's mail : the metrics described in Acta Cryst. (2014). D70, 2593-2606 "Metrics for comparison of crystallographic maps" can be applied using the stand-alone GUI python-based program (test version) which I can send you off-list if you want. Best wishes, Sacha Urzhumtsev ________________________________________ De : [email protected] de la part de Pavel Afonine Envoyé : vendredi 15 mai 2015 18:37 À : mohamed noor; [email protected] Objet : Re: [phenixbb] Map quality Hi Mohamed, this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues. Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps ... All the best, Pavel On 5/15/15 9:10 AM, mohamed noor wrote:

Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

Reply

Sign in to reply online Use email software

Murpholino Peligro

16 May 16 May

7:28 p.m.

So...Is there a tool/program/formula to get equivalent sigma levels between maps so they can be compared? 2015-05-15 11:37 GMT-05:00 Pavel Afonine :

Hi Mohamed,

this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues.

Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps

Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map

In a very nutshell,

- "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC.

All the best, Pavel

On 5/15/15 9:10 AM, mohamed noor wrote:

...
Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

Reply

Sign in to reply online Use email software

Smith Liu

17 May 17 May

8:07 a.m.

For example, can we use the phenix cut out density and then treat the phenix cut out density maps (all in mtz) from different sources (ccp4 or mtz) as comparable from the view of sigmal level? Smith At 2015-05-17 03:28:06, "Murpholino Peligro" wrote: So...Is there a tool/program/formula to get equivalent sigma levels between maps so they can be compared? 2015-05-15 11:37 GMT-05:00 Pavel Afonine : Hi Mohamed, this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues. Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map In a very nutshell, - "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC. All the best, Pavel On 5/15/15 9:10 AM, mohamed noor wrote: Dear all Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells. Thanks. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

Reply

Sign in to reply online Use email software

Pavel Afonine

19 May 19 May

5:25 a.m.

Yes: phenix.map_comparison . Pavel On 5/16/15 12:28 PM, Murpholino Peligro wrote:

So...Is there a tool/program/formula to get equivalent sigma levels between maps so they can be compared?

2015-05-15 11:37 GMT-05:00 Pavel Afonine mailto:[email protected]>:

Hi Mohamed,

this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues.

Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps

Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map

In a very nutshell,

- "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC.

All the best, Pavel

On 5/15/15 9:10 AM, mohamed noor wrote:

Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks.

_______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected] mailto:[email protected]

Reply

Sign in to reply online Use email software

Smith Liu

12:27 p.m.

Dear Pavel, Can phenix.map_comparison be used to compare CCP4 map and mtz map? How to input map files in order to use this command? Smith At 2015-05-19 13:25:28, "Pavel Afonine" wrote: Yes: phenix.map_comparison . Pavel On 5/16/15 12:28 PM, Murpholino Peligro wrote: So...Is there a tool/program/formula to get equivalent sigma levels between maps so they can be compared? 2015-05-15 11:37 GMT-05:00 Pavel Afonine : Hi Mohamed, this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues. Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map In a very nutshell, - "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC. All the best, Pavel On 5/15/15 9:10 AM, mohamed noor wrote: Dear all Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells. Thanks. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

Reply

Sign in to reply online Use email software

Pavel Afonine

4:38 p.m.

At present it uses actual map given at input, not map coefficients. Several people requested to add a functionality so that it can use map coefficients too, I will add this once I get a chance. Pavel On 5/19/15 5:27 AM, Smith Liu wrote:

Dear Pavel, Can phenix.map_comparison be used to compare CCP4 map and mtz map? How to input map files in order to use this command? Smith

At 2015-05-19 13:25:28, "Pavel Afonine" wrote:

Yes: phenix.map_comparison .

Pavel

On 5/16/15 12:28 PM, Murpholino Peligro wrote:

...
So...Is there a tool/program/formula to get equivalent sigma levels between maps so they can be compared?

2015-05-15 11:37 GMT-05:00 Pavel Afonine mailto:[email protected]>:

Hi Mohamed,

this is not a simple topic.. Two texts I suggest to have a look at are listed below I'm sure there are more). They may not give you a quick solution but perhaps will explain the issues.

Acta Cryst. (2014). D70, 2593-2606 Metrics for comparison of crystallographic maps

Acta Cryst. (2015). D71, 646-666. FEM: feature-enhanced map

In a very nutshell,

- "1 sigma (or any other x sigma) level for one map may not be the same as 1 sigma (or any other x sigma) level for another map"; - Think of "global vs local". Somehow you need to quantify map quality locally. - Map correlation (RSCC or map CC, for alternative names) may be a misleading metric is used without care: for example, two poor but similar map may give you high CC.

All the best, Pavel

On 5/15/15 9:10 AM, mohamed noor wrote:

Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks.

_______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected] mailto:[email protected]

Reply

Sign in to reply online Use email software

Jrh

15 May 15 May

6:28 p.m.

Dear Mohamed, Low completeness in low resolution shells is surely to be avoided. This should be achievable in monochromatic and white beam (Laue) measuring modes; for the latter, the most challenging, see eg http://dx.doi.org/10.1107/S0909049599006342. Anyway, from your email, you have this situation; artefacts that may arise and their nature depend on the diffraction resolution that you have and could include false peaks. For true peaks their heights will likely be degraded. Please provide more detail in your query. I would also mention that protein crystallography is entrenched in the terminology of electron density map sigmas, whereas chemical crystallography, albeit with an easier challenge of atomic resolution data to hand, present electron density maps in electrons ie an absolute scale, which is most helpful to the viewer. Best wishes, John. Emeritus Prof John R Helliwell DSc_Physics On 15 May 2015, at 17:10, mohamed noor wrote:

Dear all

Is there a single (or a few) metrics that can be used to quantitatively assess map quality instead of looking at each one in Coot? For example, I want to compare the effect of having low completeness in the low resolution shells.

Thanks. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

Reply

Sign in to reply online Use email software

3510

Age (days ago)

3514

Last active (days ago)

Download

8 comments

6 participants

tags

participants (6)

Alexandre OURJOUMTSEV
Jrh
mohamed noor
Murpholino Peligro
Pavel Afonine
Smith Liu