Re: [ccp4bb] PILATUS data collection
A couple of extra comments on top of Bob's rather comprehensive recommendations, based purely on actually looking at Pilatus data (I mean *looking*) When you are inspecting the images looking at them at 100% size is important: spots are small relative to pixels and the point spread is essentially zero. I also find it helpful to look at white spots on a black background rather than the reverse. The DECTRIS folks have an image viewer named ALBULA which works well, however ADXV also works fine (IMHO) if you tweak the settings as above. It is remarkable how big a difference it makes. The other big difference in terms of viewing the images is that (if you have low counts) you are actually at the mercy of real Poisson statistics. For example, if you have a "spot" (pixel) with 8 counts in against a background of 0 - 3 counts (say) it looks nice and clear - certainly based on experience of CCD images. And the spot is say four times the background so it must be good right? However the variance on a spot with 8 counts is 8, the variance from background subtraction may be about 3 so your spot has a maximum I/sigma of ~ 8 / sqrt(11) so about 2 and a bit. On the flip side, data from a decent crystal taken with a low dose can look almost blank at first glance (esp. with black spots on white; zoomed out) but process very nicely. Take some time to get used to the instrument. A couple of final comments: The machine has no read-out noise so fine phi slicing (and dose slicing) can only help - recording twice as many degrees with half as much dose increases the chance of getting a complete and relatively undamaged data set. All it does is take up lots of disk space. There is really no risk in doing this, unlike with a CCD where you are at war with the read-out noise. The machine also behaves completely differently to a CCD: this takes some getting used to. Take narrow oscillations as this will give better measurements of strong reflections (which I think is detailed in the papers Bob recommended.) Also *take your time* - one of the great things about these detectors is that they allow you to do continuous exposures, which can essentially double the throughput or more of data collection. Take back some of that time to use a lower dose rate and spread your photons out more evenly across reciprocal space. You can always measure more data if your sample is undamaged. Best wishes, Graeme On 7 May 2013 02:00, Robert Sweet wrote: > The seminal paper on actually how to collect data from detectors like this > and others with negligible read-out time is this one, which I strongly > recommend: > > Optimal Fine phi-slicing for Single-Photon-Counting Pixel Detectors Marcus > Mueller, Meitian Wang, and Clemens Schulze-Briese, Acta Cryst.(2012) D68, > 42-56 > > And you can pick up a copy of the paper from the RapiData web site: > http://www.px.nsls.bnl.gov/courses/papers/Mueller_ACD68_2012.pdf > > > The classic paper on data-collection strategies is this: > > Data-Collection Stragegies, Dauter, Z. Acta Cryst. (1999). D55, 1703-1717. > > Also available from the RapiData site: > > http://www.px.nsls.bnl.gov/courses/papers/dauter_strategy.pdf > > > Then there are multiple papers on damage and its impact on your data. I > suggest this one: > > Radiation damage in macromolecular crystallography: what is it and why > should we care?, E.Garman, Acta Cryst. D66, 339-351(2010). > > which you can find here: > > http://www.px.nsls.bnl.gov/courses/papers/actad-garman-2010.pdf > > With this under your belt you'll be able to decide how to collect your > phasing data. The bottom line is probably that you should go for SAD data. > Employ multiple crystals and average them together in a judicious way, > keeping only the sweeps from barely damaged x-tals. > > Good luck, > > Bob Sweet > > = > Robert M. Sweet E-Dress: sw...@bnl.gov > Group Leader, PXRR: Macromolecular ^ (that's L > Crystallography Research Resource at NSLSnot 1) > http://px.nsls.bnl.gov/ > Photon Sciences and Biosciences Dept > Office and mail, Bldg 745, a.k.a. LOB-5 > Brookhaven Nat'l Lab. Phones: > Upton, NY 11973631 344 3401 (Office) > U.S.A. 631 344 2741 (Facsimile) > = > > > On Mon, 6 May 2013, Theresa Hsu wrote: > >> Dear crystallographers >> >> Is there a good source/review/software to obtain tips for good data >> collection strategy using PILATUS detectors at synchrotron? Do we need to >> collect sweeps of high and low resolution data separately? For anomalous >> phasing (MAD), does the order of wavelengths used affect structure solution >> or limit radiation damage? >> >> Thank you. >> >> Theresa >> >
Re: [ccp4bb] how to deal 3 twin-related domains
Dear Fulvio, what did your L-Test with the unmerged (P1) data look like? In order to use twin refinement you should be sure that the structure is twinned. Did you try all possible subgroups of R3? Generally, a triplet can be treated as any other pseudo-merohedral twin, given it really is a triplet - meaning integration and scaling in the lower (real) space group, MR and twin refinement. What is the indication that MR worked? The random R value (meaning an unsolved structure) goes down from 58.5 to 50% if your structure is (pseudo-)merohedrally twinned, and you are not very much under this value. Best wishes Andrea. Am 07.05.2013 14:47, schrieb Fulvio Saccoccia: Dear ccp4 users, I have a reflections file obtained from a crystal initially scaled in H3. All the attempts to phase by Molecular Replacement were not successful and I succeeded only in P1. Based on previously solved structures from the same protein, I suspected that twinning affected crystals; so I used Twin refinement in Refmac. The program gave indication of 3 domains: Twin operator: H, K, L: Fraction = 0.335 Twin operator: H, L, -H-K-L: Fraction = 0.332 Twin operator: H, -H-K-L, K: Fraction = 0.333 With P1 and twin refinement I obtained a reasonable initial value of FOM (0.50) but R/Rfree still remain at very large values (0.46/0.49). I am not sure about twinning and data handling in presence of 3 twin domains and I will appreciate a lot if someone will suggest me how to deal out with these data. Thanks Fulvio -- Dr. Andrea Thorn Department of Haematology, University of Cambridge Cambridge Institute for Medical Research Wellcome Trust/MRC Building Hills Road Cambridge CB2 0XY, U.K. andrea.th...@web.de http://shelx.uni-ac.gwdg.de/~athorn Tel: + 44 1223 763234
[ccp4bb] Postdoctoral opening - Univ Liverpool
Dear colleagues, We are offering a 3-yr postdoctoral position on the structural study of muscle proteins linked to heart disease. The project involves the attractive integration of X-ray crystallography, bioinformatics and protein complexation methods at large scale. Full details on the position, our laboratory infrastructure and how to apply can be found at: http://www.liv.ac.uk/working/job_vacancies/research/r-582983/ Please direct any informal enquiries to: olga.may...@liv.ac.uk (Deadline: Friday 7th June, 2013) Olga Mayans Inst. Integrative Biology University of Liverpool
Re: [ccp4bb] PILATUS data collection
Graeme and Bob, Wow... It's great to learn from actual experiences. Thanks much for this write up. If this were stackoverflow, +1. F On May 8, 2013, at 12:32 AM, Graeme Winter wrote: > A couple of extra comments on top of Bob's rather comprehensive > recommendations, based purely on actually looking at Pilatus data (I > mean *looking*) > > When you are inspecting the images looking at them at 100% size is > important: spots are small relative to pixels and the point spread is > essentially zero. I also find it helpful to look at white spots on a > black background rather than the reverse. > > The DECTRIS folks have an image viewer named ALBULA which works well, > however ADXV also works fine (IMHO) if you tweak the settings as > above. It is remarkable how big a difference it makes. > > The other big difference in terms of viewing the images is that (if > you have low counts) you are actually at the mercy of real Poisson > statistics. For example, if you have a "spot" (pixel) with 8 counts in > against a background of 0 - 3 counts (say) it looks nice and clear - > certainly based on experience of CCD images. And the spot is say four > times the background so it must be good right? > > However the variance on a spot with 8 counts is 8, the variance from > background subtraction may be about 3 so your spot has a maximum > I/sigma of ~ 8 / sqrt(11) so about 2 and a bit. On the flip side, data > from a decent crystal taken with a low dose can look almost blank at > first glance (esp. with black spots on white; zoomed out) but process > very nicely. Take some time to get used to the instrument. > > A couple of final comments: > > The machine has no read-out noise so fine phi slicing (and dose > slicing) can only help - recording twice as many degrees with half as > much dose increases the chance of getting a complete and relatively > undamaged data set. All it does is take up lots of disk space. There > is really no risk in doing this, unlike with a CCD where you are at > war with the read-out noise. > > The machine also behaves completely differently to a CCD: this takes > some getting used to. Take narrow oscillations as this will give > better measurements of strong reflections (which I think is detailed > in the papers Bob recommended.) Also *take your time* - one of the > great things about these detectors is that they allow you to do > continuous exposures, which can essentially double the throughput or > more of data collection. Take back some of that time to use a lower > dose rate and spread your photons out more evenly across reciprocal > space. You can always measure more data if your sample is undamaged. > > Best wishes, > > Graeme > > On 7 May 2013 02:00, Robert Sweet wrote: >> The seminal paper on actually how to collect data from detectors like this >> and others with negligible read-out time is this one, which I strongly >> recommend: >> >> Optimal Fine phi-slicing for Single-Photon-Counting Pixel Detectors Marcus >> Mueller, Meitian Wang, and Clemens Schulze-Briese, Acta Cryst.(2012) D68, >> 42-56 >> >> And you can pick up a copy of the paper from the RapiData web site: >> http://www.px.nsls.bnl.gov/courses/papers/Mueller_ACD68_2012.pdf >> >> >> The classic paper on data-collection strategies is this: >> >> Data-Collection Stragegies, Dauter, Z. Acta Cryst. (1999). D55, 1703-1717. >> >> Also available from the RapiData site: >> >> http://www.px.nsls.bnl.gov/courses/papers/dauter_strategy.pdf >> >> >> Then there are multiple papers on damage and its impact on your data. I >> suggest this one: >> >> Radiation damage in macromolecular crystallography: what is it and why >> should we care?, E.Garman, Acta Cryst. D66, 339-351(2010). >> >> which you can find here: >> >> http://www.px.nsls.bnl.gov/courses/papers/actad-garman-2010.pdf >> >> With this under your belt you'll be able to decide how to collect your >> phasing data. The bottom line is probably that you should go for SAD data. >> Employ multiple crystals and average them together in a judicious way, >> keeping only the sweeps from barely damaged x-tals. >> >> Good luck, >> >> Bob Sweet >> >> = >>Robert M. Sweet E-Dress: sw...@bnl.gov >>Group Leader, PXRR: Macromolecular ^ (that's L >> Crystallography Research Resource at NSLSnot 1) >> http://px.nsls.bnl.gov/ >>Photon Sciences and Biosciences Dept >>Office and mail, Bldg 745, a.k.a. LOB-5 >>Brookhaven Nat'l Lab. Phones: >>Upton, NY 11973631 344 3401 (Office) >>U.S.A. 631 344 2741 (Facsimile) >> = >> >> >> On Mon, 6 May 2013, Theresa Hsu wrote: >> >>> Dear crystallographers >>> >>> Is there a good source/review/software to obtain tips for go
[ccp4bb] Off-topic (SPR)
Dear pals, Sorry for my off-topic question but its been a while I am challenging to get a 15 kDa protein coated on CM5 chip. Calculated PI of protein is 5.9 & I tried Acetate buffer pH5.2 and 3.8 for coating under normal procedure of Biacore. After immobilization I saw this message: response level decreased! This happened for me several times. Does anyone have any idea what's going on? Cheers, Jahan
Re: [ccp4bb] Off-topic (SPR)
Hi Jahan, Maybe you are using a new model of BiaCore. Our old BiaCore X never cares to tell us how to do our experiment. However if your RU did go below that of the empty chip, then maybe you really didn't get anything conjugated. The pH 3.8 condition is not good for the amine coupling reaction. The reaction will be very inefficient when pH drops below 5. You can try conjugating at higher pH, such as 6, 5.5 or 5. If you can not get protein adsorbed to the surface by electrostatic attraction, then simply increase the protein concentration - 0.05, 0.1mg/mL, or as much as you can afford to use, and extend the time of reaction (you can slow down or even stop the flow of protein to get some "incubation"). You should see at least 100 or 200 RU increase after a successful coupling for your 15kD protein. How many Lys do you have in the 15kDa protein? If you have very few Lys, and the protein is made in E coli, you can add a short poly Lys tail to either ends of the protein. This also has an additional advantage if some of the Lys residues are located in the binding surface. Zhijie From: Jahan Alikhajeh Sent: Wednesday, May 08, 2013 3:08 PM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] Off-topic (SPR) Dear pals, Sorry for my off-topic question but its been a while I am challenging to get a 15 kDa protein coated on CM5 chip. Calculated PI of protein is 5.9 & I tried Acetate buffer pH5.2 and 3.8 for coating under normal procedure of Biacore. After immobilization I saw this message: response level decreased! This happened for me several times. Does anyone have any idea what's going on? Cheers, Jahan
[ccp4bb] question about CCP4 scripts
Hi All,
I am trying to write a shell script to streamline a few steps, one of which
is Unique, see below. As you can see, this program requires symmetry and
cell parameters. In CCP4 GUI Scalepack2mtz, these info are automatically
extracted from .sca file (first two lines). But I don't know if there is a
way to do this in script, so I don't need to type these values for each
dataset. Thank you in advance for your help.
#!/bin/sh
# unique.exam
#
# runnnable test script for the program "unique" - this will use this
# program to generate a reflection list containing null values.
#
set -e
unique hklout ${CCP4_SCR}/unique_out.mtz << eof
labout F=F SIGF=SIGF
symmetry p43212
resolution 1.6
cell 78.1 78.1 55.2 90.0 90.0 90.0
eof
--
Best regards,
Joe
Re: [ccp4bb] question about CCP4 scripts
Hi Joe,
this line of code for csh/tcsh is awfully rickety-raggedy, it does the trick
though.
Jan
setenv spacegroup P3121
setenv noresidues 350
setenv lambda 1.5418
set unitcell = `head -3 ../hkl/output.sca | tail -1 | cut -b 4-60`
scalepack2mtz \
hklin ../hkl/output.sca \
hklout ./${dataset}_import.mtz \
<< eof > ${dataset}_scalepack2mtz.log
symmetry $spacegroup
cell $unitcell
wave $lambda
end
eof
--
Jan Abendroth
Emerald BioStructures
Seattle / Bainbridge Island WA, USA
home: Jan.Abendroth_at_gmail.com
work: JAbendroth_at_embios.com
http://www.emeraldbiostructures.com
On May 8, 2013, at 8:37 PM, Joe Chen wrote:
> Hi All,
>
>
> I am trying to write a shell script to streamline a few steps, one of which
> is Unique, see below. As you can see, this program requires symmetry and
> cell parameters. In CCP4 GUI Scalepack2mtz, these info are automatically
> extracted from .sca file (first two lines). But I don't know if there is a
> way to do this in script, so I don't need to type these values for each
> dataset. Thank you in advance for your help.
>
>
> #!/bin/sh
> # unique.exam
> #
> # runnnable test script for the program "unique" - this will use this
> # program to generate a reflection list containing null values.
> #
>
> set -e
>
> unique hklout ${CCP4_SCR}/unique_out.mtz << eof
> labout F=F SIGF=SIGF
> symmetry p43212
> resolution 1.6
> cell 78.1 78.1 55.2 90.0 90.0 90.0
> eof
>
>
> --
> Best regards,
>
> Joe
