Project

General

Profile

MSI raster question

Added by Anonymous over 18 years ago

I'm wondering about the feasibility of using the MSI raster application with cold lacey grids. We may have a specimen problem with a protein that tends to stick to the carbon and not go into the holes on quantifoil grids; so maybe it would be possible to use lacey grids instead? In terms of targeting this may mean just taking pictures of an entire square and discarding the images of areas with carbon afterwards. However, there would be problems with getting suitable areas for focussing on a lacey grid - I think at high magnification if a focus target ended up in a hole it would be too low contrast? Also, I think the Leginon manual suggests that suitable applications for MSI raster are pure carbon grids and/ or negatively stained grids rather than lacey grids?

William


Replies (1)

- Added by Jim Pulokas over 18 years ago

I'm not too familiar with using lacey grids, but you may find both the raster node and the regular hole finder node can be set up to work in a lot of non standard situations. For instance, Anchi has been using the hole finder to locate squares on a low mag image of the grid. The squares are actually the "holes" that are being found.

With either targeting node, you have the ability to filter out a certain ice thickness from the final target list. This is probably still going to be useful on your lacey grids. The problem is choosing either the raster finder or the hole finder. For your high mag target finder, you will probably want to use the hole finder (JAHC finder) because it has the option "Use focus template thickness". this will allow you to set up a focus target template with many potential focus positions and then it will pick one that has the proper ice thickness (in this case, carbon thickness) to focus on, while rejecting the improper ice thickness (holes). The trick is setting up all the other parameters to get your main exposure targets selected automatically. You will need to design an appropriate hole template image that produces a lot of potential blobs, then set a 1 blob limit so that it only finds the one closest to the center.

For low mag targeting on the squares, you may just want to do a raster over the whole square, since your high mag targeting will do most of the work in picking the best locations. You will want to create a raster over most of the image, then use the ice thickness thresholds to reject targets that are on grid bars. Since you are covering selecting targets over the whole square, you then leave it up to the high mag targeting to determine if there is actually anything useful at that location.

So this may require a hybrid application with a hole finder and a raster finder. This is not too difficult because you can start with either MSI Raster or the MSI with hole finder, and just modify one of the nodes. It may be easiest to do this by saving the XML file, making the changes, then importing it back in to the database.

Jim

Jim

    (1-1/1)