Hi,

Replies to your specific queries are interspersed with your text below.

> Message du 19/12/09 05:19
> De : "Jinzhong Lin" 

> A : [log in to unmask]
> Copie à : 
> Objet : [ccp4bb] merge data from multiple isomorphous crystals.
> 
> 
> Dear all,
> I have a crystal that can diffract to 3.0A for the first few frames at
> synchrotron, after procession only up to 3.6A data are useful. The
> structure has been solved by MAD. In order to get a 3.0A dataset, I am
> going to collect the first few images from dozens of isomorphous
> crystals and merge them together. Because i have no experience like this
> before, i am here asking for some help or suggestions. Especially I have
> some specific questions:
> 1. Is it possible to extend the resolution of the structure to 3.0A by
> this method?

At the time of data collection in X-ray cappillaries and room temperature or close to 0 deg, this is what one used to do. All the time, except for the few 'rocks' that lasted for ever.

BTW, you did not mention: are you using data collection with cryo-cooled crystals? If yes, it may be worthwhile to check if you get the same behaviour with crystals mounted in cappillaries, you could be lucky!

> 2. At least how many frames should i collect for a single crystal for
> data procession?

I can only speak about the processing program I have used recently: XDS, I get acceptable results when I have at least 5 to 10 frames. But virus people used to be able to collect only a single image per crystal (at the time of data collection in cappillaries, this could also be the case now for crystals of viruses that are sufficiently dangerous to force enclosure in cappillaries).

> 3. Since the crystal can grow to a large dimension (0.3x0.3x0.4mm), can
> i collect the dataset at different positions of the crystal so that the
> a single crystal can be re-used for several times thus to reduce the
> number of total crystals needed.

Yes, this is done quite often. You still have crystal degradation that extends to areas of the crystal that haven't been exposed though.

> 4. How many frames in total should be enough for final data procession
> and scale.

This depends on the space-group.

> 5. which programs(hkl, xds, ...) are the best for merging the dataset.

Question of taste, all programs do equally well if you know how to use them properly. I hear lots of good things about SAINT but I haven't laid my hands on this software suite yet. In particular a talk by Bram Schierbeek at ESRF last week, with one structure solved using Molybdenum radiation and sulphur SAD.

> 6. is there anything i need to pay particular attention to during data
> collection and procession.

Not all synchrotron beam lines have a set up that allow you to collect data in cappillaries should that be needed. Otherwise, the usual applies (stay away from synchrotrons where the canteen is not so good :-)

> 7. Some tricks, tips or references for such a job.

Check older references. I could always provide a list of my own publications but that would not be fair.

> 
> I appreciate any responses.
> 
> Best Regards,
> lithin
> 
> 

Fred.