[phenixbb] HKL2000 behaved weirdly

Mengbin Chen mengbinc at sas.upenn.edu
Tue Aug 20 11:51:06 PDT 2013

Sorry for being late in replying, but thank you all for your suggestions!
The resolution is ~3 angstroms, so it was quite easy to pick more than 100
peaks for indexing. So I tried to index ~5-6 images at the same time as you
suggested, the Bravais Lattice window appeared (and I reset the beam center
too), but the unit cell dimensions were ~10 angstroms along one axis, which
was not appropriate since the protein is around 40 kDa. I tried to manually
pick peaks for indexing, still cell dimensions were too small to be
properly assigned. I feel the crystal has got severe problems of multiple
lattices and high mosaicity at the same time, which made it hard to index.
I am still playing around with the dataset to see if I can get any luck
with it, so I'll let you know if I can work this out.


On Fri, Aug 16, 2013 at 12:18 PM, Noinaj, Nicholas (NIH/NIDDK) [F] <
noinajn at niddk.nih.gov> wrote:

> Mengbin,
> This happens in HKL2000 with tough datasets.  The data can sometime appear
> to be very nice, possibly some high mosaicity, but still not process
> properly.  I have seen this many times.  As mentioned already, the next
> step is to start trying different things to get the data to process.
>  WIthout knowing more about resolution, etc, here is a few things to try.
> 1- Check and the double check the beam center.  Then, check it again!
>  With troublesome datasets, this is EXTREMELY important.  In HKL2000,
> everytime you hit Abort Refinement, you have to reset the beamcenter again.
>  SO monitor the beam center and everytime it gets too far away from the
> actual numbers, reset it.
> 2- you want to try to have roughly 100 peaks for indexing at minimum (if
> possible), so change the resolution limits to 15 ang - 4 ang, then retry
> indexing.  IF this does not work, try 15-5ang, then 12-5, then 10-5, etc.
> 3- If #2 doesn't work, then you can alway try indexing using multiple
> frames, esp if you don't have a enough peaks on the first frame.  To do
> this, please see the HKL2000 manual, or let me know, i can send you some
> screenshots or give you a call to guide you via phone.  here, you can use
> 5-10 frames for indexing if needed.
> 4- as a last option, you can always manually select your peaks. i have had
> this work nicely in the past, but again, last option.
> 5- you could try other data processing programs too such as MoSFLM, XDS,
> Xia2, etc.
> If you still don't get it working, i am more than happy to take a quick
> look.  Just send me one of your frames to analyse.  Of course, everything
> will be strictly confidential.  Thanks in advance and good luck!
> Cheers,
> Nick
> ________________________________
> From: Mengbin Chen [mengbinc at sas.upenn.edu]
> Sent: Thursday, August 15, 2013 3:12 PM
> To: PHENIX user mailing list
> Subject: [phenixbb] HKL2000 behaved weirdly
> Dear All,
> I used HKL2000 to index my data, but weird things happened: after I hit
> peak search, the data were selected to index, and I chose primitive
> triclinic to continue. After I hit refine, the circles for peak search
> disappeared, and if I hit "Bravais Lattice", the table would not show up,
> which meant that I could not proceed to refine in the potentially correct
> Bravais Lattice. I don't know if anyone else has encountered this problem
> before, so any suggestions would be greatly appreciated!
> Thank you in advance,
> Mengbin
> --
> Mengbin Chen
> Department of Chemistry
> University of Pennsylvania

Mengbin Chen
Department of Chemistry
University of Pennsylvania
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20130820/55ebd3ce/attachment-0001.htm>

More information about the phenixbb mailing list