General work flow for protein SAXS data processing

Interactive and automated data processing tools (PRIMUS, GNOM, AUTORG).
Scattering from simple bodies (BODIES), peak analysis (PEAK), data plotting (SASPLOT) etc.
Post Reply
Message
Author
Donghyuk Shin
Member
Posts: 9
Joined: 2013.04.12 13:58

General work flow for protein SAXS data processing

#1 Post by Donghyuk Shin » 2013.04.15 17:20

I have just entered SAXS world, and I have no-idea for manipulating my data.
Until now, I have searched a lot, read a lot, tried a lot, and got some idea for handling my data.
Still, however, I do not know that whether my strategies are correct or not. I need some good advisor for learning SAXS, please help me.

Here are some questions along with the story.

[Q1] Here are my strategies for data processing, and Is this scheme is OK for general biosaxs data processing? or I miss something? Do I have to choose another program for some case? Please help me.

1)For single peptide chain,
I did PRIMUS-GNOM-DAMMIF-SUPCOMB as an order.

2)For oligomeric protein case,
I did PRIMUS-GNOM-DAMMIF-SASREF-SUPCOMB as an order. (Did CRYSOL for my pdb model, and could ger amplitude file)

---continue
In details, When I did PRIMUS-GNOM I just click the data on primus, and click the distance distribution, and I could get *.out file as results.
[Q2] Is it OK that I just use raw data without considering the range of it? when I did radius of gyration analysis, It gave me the range of good fitting, and I think I should use that range for *.out file. I do not know whether PRIMUS can automatically generate the *.out file for good fitting or I must specify the range for it.

[Q3] If I should specify the range, which range should I select and how can I generate the trimmed data file from raw data. (If there is some easy way to generate trimmed data, please let me know, for now I just deleting each line by editing data using vi editor on terminal.)

---continue
After generating *.out file, I used it as an input file for DAMMIF. I typed the default values, and got dammit-1.pdb file (I know that I can change the variables, and name). I also got *.fir, *.fit file, and I could find chi value from the first line of *.fir file.
[Q4] How that chi should be? What is the generally acceptable chi, good chi, and bad chi for biosaxs?

---continue
I checked my chi were 1.3, and I accept it myself. I go through the next step, and I did SUPCOMB to align the dammit-1.pdb to crystal_structure_model.pdb
I got the good fitted projection, and I saw that NSD value from aligned pdb is around 0.3. I wrote down about this at saxier forum. I got answer that I should run SUPCOMB 10-20 cycles to fit in with good value of NSD.

[Q5] How the NSD should be? The reasonable, nice, and bad case of NSD. I knew that theoretically it should be 0, and over 1 is not good. However, I wants to know that in real case. I think you are the only one who can answer this question from lots of experiences.

---continue
[Q6]I did not know that I should run 10-20 cycles of SUPCOMB before getting answer from forum, Is their any refinement steps for each level of data processing?

I do need kind help, and advice. Please help me.

sachrysts
Active member
Posts: 70
Joined: 2011.09.07 15:53
Location: Stanford

Re: General work flow for protein SAXS data processing

#2 Post by sachrysts » 2013.04.16 15:30

Donghyuk Shin wrote: In details, When I did PRIMUS-GNOM I just click the data on primus, and click the distance distribution, and I could get *.out file as results.
[Q2] Is it OK that I just use raw data without considering the range of it? when I did radius of gyration analysis, It gave me the range of good fitting, and I think I should use that range for *.out file. I do not know whether PRIMUS can automatically generate the *.out file for good fitting or I must specify the range for it.
1) Do the radius of gyration (i.e. autorg) analysis. It will give you an estimate for the Rg and the range where it thinks the "good" data starts.
2) Primus will take this range and the Rg for the Distance distribution (i.e. gnom) analysis.
Donghyuk Shin wrote: [Q3] If I should specify the range, which range should I select and how can I generate the trimmed data file from raw data. (If there is some easy way to generate trimmed data, please let me know, for now I just deleting each line by editing data using vi editor on terminal.)
You can use Primus to select the range and save via right-click on the filename.
Donghyuk Shin wrote: [Q4] How that chi should be? What is the generally acceptable chi, good chi, and bad chi for biosaxs?
Did you use the search function of this forum or google with terms like, good chi value saxs?
Donghyuk Shin wrote: I checked my chi were 1.3, and I accept it myself. I go through the next step, and I did SUPCOMB to align the dammit-1.pdb to crystal_structure_model.pdb
I got the good fitted projection, and I saw that NSD value from aligned pdb is around 0.3. I wrote down about this at saxier forum. I got answer that I should run SUPCOMB 10-20 cycles to fit in with good value of NSD.
You should not run supcomb 10 times but dammif 10-20 times. Read carefully.
Donghyuk Shin wrote: [Q5] How the NSD should be? The reasonable, nice, and bad case of NSD. I knew that theoretically it should be 0, and over 1 is not good. However, I wants to know that in real case. I think you are the only one who can answer this question from lots of experiences.
Again use the forum's search function on NSD values and read the manual, please.

User avatar
AL
Administrator
Posts: 695
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Primary SAX data reduction and analysis steps

#3 Post by AL » 2013.04.17 16:05

The first SAXS data processing steps topic in FAQ might be a good starting point.

Donghyuk Shin
Member
Posts: 9
Joined: 2013.04.12 13:58

Re: General work flow for protein SAXS data processing

#4 Post by Donghyuk Shin » 2013.04.18 06:17

Thank you both sachrysts, and AL. :)

It seems like I should read and study more and carefully!
Thank you for all information and advice!

Post Reply