2 Wonderings: adding H atoms and overcoming overfitting problem

older
optimizing target weights in phenix

Young-Jin Cho

16 Feb 2009 16 Feb '09

4:34 p.m.

Hi, my current wonderings are as below: First, how can I add H atoms into nucleotide ligands in pdb file? I did 'phenix.reduce model.pdb > model_h.pdb', but as I see it didn't add H's of sugar (1', 2', 3' 4' and 5'). Once I had this kind of problem before, I was suggested to use phenix.ready_set command but seemingly this does not work any longer with last version that I am using now. Also, is it possible to add H into phosphate group? The second very general question is about getting over overfitting problem. My phenix runs gave me too much difference between R and Rfree about .7 w/o adding H atoms. I assume it came from water molecules because it(huge gap) just starts to happen after 1_bss stage. If anybody can suggest a good way to get over this problem, you will save my tremendous times and useless efforts. Big thanks in advance, Young-Jin

Show replies by date

junfeng liu

16 Feb 16 Feb

5:01 p.m.

New subject: 2 Wonderings: adding H atoms and overcoming overfitting problem

Hi Young-Jin, For the second one you can try lower the scale factor of wxu and wxc from default (1 and 0.5) . Is it the first round to run phenix? If not better to go back and rerun it again. See the manual for details: http://www.phenix-online.org/documentation/refinement.htm And for the first one |I have no experience on that and no idea at all. Hope other guys will help you for that. Good luck! leo Young-Jin Cho wrote:

...

Hi, my current wonderings are as below: First, how can I add H atoms into nucleotide ligands in pdb file? I did 'phenix.reduce model.pdb > model_h.pdb', but as I see it didn't add H's of sugar (1', 2', 3' 4' and 5'). Once I had this kind of problem before, I was suggested to use phenix.ready_set command but seemingly this does not work any longer with last version that I am using now. Also, is it possible to add H into phosphate group?

The second very general question is about getting over overfitting problem. My phenix runs gave me too much difference between R and Rfree about .7 w/o adding H atoms. I assume it came from water molecules because it(huge gap) just starts to happen after 1_bss stage. If anybody can suggest a good way to get over this problem, you will save my tremendous times and useless efforts.

Big thanks in advance,

Young-Jin

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

Nigel W Moriarty

17 Feb 17 Feb

4:29 p.m.

New subject: 2 Wonderings: adding H atoms and overcoming overfitting problem

Young-Jin Which version of PHENIX are you using? If you send your files directly to me I'll determine what is happening. Nigel On 2/16/09 8:34 AM, Young-Jin Cho wrote:

...

Hi, my current wonderings are as below: First, how can I add H atoms into nucleotide ligands in pdb file? I did 'phenix.reduce model.pdb > model_h.pdb', but as I see it didn't add H's of sugar (1', 2', 3' 4' and 5'). Once I had this kind of problem before, I was suggested to use phenix.ready_set command but seemingly this does not work any longer with last version that I am using now. Also, is it possible to add H into phosphate group?

The second very general question is about getting over overfitting problem. My phenix runs gave me too much difference between R and Rfree about .7 w/o adding H atoms. I assume it came from water molecules because it(huge gap) just starts to happen after 1_bss stage. If anybody can suggest a good way to get over this problem, you will save my tremendous times and useless efforts.

Big thanks in advance,

Young-Jin

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

Young-Jin Cho

19 Feb 19 Feb

5:24 p.m.

New subject: 2 Wonderings: adding H atoms and overcoming overfitting problem

Nigel, Have you had a chance to look at my pdb files? Anyway, I was suggested to use molprobity website and I could get hydrogen added pdb file from there. I am inclined to think it is another bug of phenix though, which I thought was solved from new version. Just let me know what the problem is. Then have a good day! Young-Jin ----- Original Message ----- From: "Nigel W Moriarty" To: "PHENIX user mailing list" Sent: Tuesday, February 17, 2009 11:29:05 AM GMT -05:00 US/Canada Eastern Subject: Re: [phenixbb] 2 Wonderings: adding H atoms and overcoming overfitting problem Young-Jin Which version of PHENIX are you using? If you send your files directly to me I'll determine what is happening. Nigel On 2/16/09 8:34 AM, Young-Jin Cho wrote:

...

Hi, my current wonderings are as below: First, how can I add H atoms into nucleotide ligands in pdb file? I did 'phenix.reduce model.pdb > model_h.pdb', but as I see it didn't add H's of sugar (1', 2', 3' 4' and 5'). Once I had this kind of problem before, I was suggested to use phenix.ready_set command but seemingly this does not work any longer with last version that I am using now. Also, is it possible to add H into phosphate group?

The second very general question is about getting over overfitting problem. My phenix runs gave me too much difference between R and Rfree about .7 w/o adding H atoms. I assume it came from water molecules because it(huge gap) just starts to happen after 1_bss stage. If anybody can suggest a good way to get over this problem, you will save my tremendous times and useless efforts.

Big thanks in advance,

Young-Jin

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

-- Nigel W. Moriarty Building 64R0246B, Physical Biosciences Division Lawrence Berkeley National Laboratory Berkeley, CA 94720-8235 Phone : 510-486-5709 Fax : 510-486-5909 Email : [email protected] Web : CCI.LBL.gov _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb -- Postdoctoral Fellow, Kern Lab Biochemistry Department/HHMI Brandeis University

Nigel W Moriarty

5:57 p.m.

New subject: 2 Wonderings: adding H atoms and overcoming overfitting problem

Young-Jin I'm sorry for the delay. To add hydrogens you should use ReadySet! which uses Reduce for the protein and RNA/DNA portions of your model. For other molecules such as ligands it uses eLBOW. phenix.ready_set model.pdb will generate a file model.updated.pdb which contains the hydrogens. ReadySet! also performs some other tasks to prepare your model for refinement. Nigel On 2/19/09 9:24 AM, Young-Jin Cho wrote:

...

Nigel, Have you had a chance to look at my pdb files? Anyway, I was suggested to use molprobity website and I could get hydrogen added pdb file from there. I am inclined to think it is another bug of phenix though, which I thought was solved from new version. Just let me know what the problem is.

Then have a good day!

Young-Jin

----- Original Message ----- From: "Nigel W Moriarty" To: "PHENIX user mailing list" Sent: Tuesday, February 17, 2009 11:29:05 AM GMT -05:00 US/Canada Eastern Subject: Re: [pdbphenixbb] 2 Wonderings: adding H atoms and overcoming overfitting problem

Young-Jin

Which version of PHENIX are you using? If you send your files directly to me I'll determine what is happening.

Nigel

On 2/16/09 8:34 AM, Young-Jin Cho wrote:

...
Hi, my current wonderings are as below: First, how can I add H atoms into nucleotide ligands in pdb file? I did 'phenix.reduce model.pdb > model_h.pdb', but as I see it didn't add H's of sugar (1', 2', 3' 4' and 5'). Once I had this kind of problem before, I was suggested to use phenix.ready_set command but seemingly this does not work any longer with last version that I am using now. Also, is it possible to add H into phosphate group?

The second very general question is about getting over overfitting problem. My phenix runs gave me too much difference between R and Rfree about .7 w/o adding H atoms. I assume it came from water molecules because it(huge gap) just starts to happen after 1_bss stage. If anybody can suggest a good way to get over this problem, you will save my tremendous times and useless efforts.

Big thanks in advance,

Young-Jin

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

5626

Age (days ago)

5629

Last active (days ago)

List overview

Download

4 comments

3 participants

participants (3)

junfeng liu
Nigel W Moriarty
Young-Jin Cho