resolution effects in Cryson via the fourth column

Calculation of SAXS and SANS profiles (CRYSOL, CRYSON), superposition of models (SUPCOMB, DAMAVER, DAMCLUST), database DARA
Post Reply
Message
Author
User avatar
anlarsen
Member
Posts: 11
Joined: 2017.03.06 16:37
Contact:

resolution effects in Cryson via the fourth column

#1 Post by anlarsen » 2019.01.24 10:06

Dear ATSAS (cryson) developers,

most SANS beamlines nowadays provide the resolution effect via an uncertainty of q, in the fourth column of data. However, as of now it is not possible to use this information to take resolution effects into account in Cryson. I know, that one can provide a resolution file. However, using the fourth column is easier and is in many cases also more precise.

will you provide this option? coding-wise it should be relatively straight-forward.

the option is provided in Pepsi-SANS (https://team.inria.fr/nano-d/software/pepsi-sans/), but Pepsi-SANS lacks an option to perdeuterate separate chains.
the option is also provided in CaPP (https://github.com/Niels-Bohr-Institute ... ophys/CaPP), but fitting in CaPP is not well-suited for batch-mode, as it runs in a python GUI.

Best regards,
Andreas

User avatar
AL
Administrator
Posts: 670
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

fourth column in SANS data

#2 Post by AL » 2019.01.25 16:59

anlarsen wrote:
2019.01.24 10:06
most SANS beamlines nowadays provide the resolution effect via an uncertainty of q, in the fourth column of data.
Could you please point us to some documentation on how the fourth column should be interpreted? So far we've seen SANS data in four-column format only from KWS1 (FRM2, Munich).

User avatar
anlarsen
Member
Posts: 11
Joined: 2017.03.06 16:37
Contact:

Re: resolution effects in Cryson via the fourth column

#3 Post by anlarsen » 2019.02.06 18:56

Sure I can.

Several SANS beamlines provide a fourth column in data, e.g. these:
QUAKKA@ANSTO
D22@ILL

... and probable many more. The fourth column provides an uncertainty on the estimated q-value (1 sigma, assuming normal distributions).

Documented e.g. here:
https://www.ncnr.nist.gov/staff/hammoud ... ter_15.pdf

that is, the value of I at a given q_i is calculated as the integral of I(q) over a range of q values, weighted with the normal distribution with mean q_i, and standard deviation s_i. Here s_i denotes the spread of the i'th q (sigma of q). s_i is (often) given in the fourth column of data.

Post Reply