<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<STYLE type=text/css> P, UL, OL, DL, DIR, MENU, PRE { margin: 0 auto;} </STYLE>
<META name=GENERATOR content="MSHTML 11.00.9600.18315"></HEAD>
<BODY leftMargin=1 rightMargin=1 topMargin=1><FONT color=black size=2 face="Segoe UI">
<DIV>To Michael's list I would also suggest that you consider reducing your sample volume and/or spatial resolution.</DIV>
<DIV> </DIV>
<DIV>I don't know what spatial resolution you require, but for cubic 1um voxels there are 1e9 voxels/uL and then consider you probably have 2 bytes/voxel times 4 channels or ~8 GB/uL (per timepoint). If you can fully utilize the camera speed (a big assumption) you can get 800 MB/sec, so you are optimistically ~10 sec/uL overall acquisition rate. So to sample 100s of uL multiple times per minute is unrealistic at 1um sort of spatial resolution. And that's even sweeping under the rug the fact that you'll need to acquire more data to do the post-processing resulting in isotropic resolution. And it also assumes you can keep acquiring this sort of data for several days without running out of data storage.</DIV>
<DIV> </DIV>
<DIV>I would suggest you try to avoid imaging the entirety of your sample, and instead figure out where the cells are and acquire just those regions at every timepoint. You could ideally adjust tweak the position of the acquisition regions based on automated analysis of the position of the cells at the last time point.</DIV>
<DIV> </DIV>
<DIV>The diSPIM geometry avoids having to rotate the sample for isotropic resolution, but adds another camera. The scattering problem is also 2x worse in diSPIM geometry because you can't rotate the sample 180 degrees. But on the other hand the open-top geometry may be more convenient for your sample.</DIV>
<DIV> </DIV>
<DIV>Jon</DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV></DIV>
<DIV><FONT color=red>> Hi Mel,<BR>><BR>> that’s a long list of requirements. Here are some thoughts:<BR>><BR>> - test your samples on existing light sheet setups<BR>> - also test/consider a diSPIM-like configuration<BR>> - 2-Photon is beyond reach for $100k - and hard to do for 3-4 channels<BR>> - a fiber-coupled laser launch with 4 lines will be expensive - you might<BR>> want to start with less lines and upgrade later<BR>> - a feasible detection combination would be a single sCMOS camera plus a<BR>> filter wheel<BR>><BR>> Best,<BR>> Michael<BR>><BR>> > On Jun 4, 2016, at 7:45 PM, Menelaos Symeonides <msymeoni@UVM.EDU> wrote:<BR>> ><BR>> > A little correction... the velocity of the cells is 15 µm/min, not /sec,<BR>
> meaning that the desired imaging speed would be on the order of 2-3<BR>> volumes/min, which should be a bit more manageable!<BR>> ><BR>> > Mel<BR>> ><BR>> > -------- Original message --------<BR>> > From: Menelaos Symeonides <msymeoni@uvm.edu><BR>> > Date: 6/2/16 7:48 AM (GMT-05:00)<BR>> > To: <A href="mailto:openspim@openspim.org"><FONT color=#0000ff>openspim@openspim.org</FONT></A><BR>> > Subject: [OpenSPIM] Planning an appropriate SPIM build for my experiments<BR>> ><BR>> > Hello OpenSPIM,<BR>> ><BR>> > I've been following this list and reading papers in the field for about<BR>> > a year. In our lab, we've come to a point where we need to carry out<BR>
> > long term fast nontoxic imaging of cell cultures that is only possible<BR>> > with SPIM, and have just obtained funding to build our own microscope to<BR>> > do these experiments. I was hoping to pick the brains of this mailing<BR>> > list to hopefully design an appropriate setup, perhaps based on the<BR>> > OpenSPIM platform, so that we can carry out these experiments.<BR>> ><BR>> > In general, the experiment involves imaging HIV-infected T cells<BR>> > embedded in a 3D collagen hydrogel as they migrate through the ECM and<BR>> > spread the infection to other cells and do various other things related<BR>> > to the infection, e.g. bystander cell killing. Here are some of the<BR>> > considerations:<BR>> ><BR>
> > 1. We have fluorescently-tagged viruses in the FITC channel, a live cell<BR>> > apoptosis sensor in the iRFP channel, and will also need to somehow<BR>> > label nuclei and perhaps also membranes. That would be a maximum of 4<BR>> > channels, with 3 channels being "acceptable".<BR>> ><BR>> > 2. The cells will need to be imaged for 2 to 3 days, with a time<BR>> > resolution on the order of 2 Hz (for the full volume). T cells migrate<BR>> > at about 15 µm/sec, and we would like sufficient time resolution to<BR>> > capture each cell at least twice during each body length movement (body<BR>> > length is about 10 µm, hence needing around 2 Hz).<BR>> ><BR>> > 3. The entire volume (or close to that) of the hydrogel will need to be<BR>
> > imaged at each timepoint. The volume of the hydrogel is up to our<BR>> > discretion, but generally it is on the order of 150 to 300 µL.<BR>> ><BR>> > 4. Spatial resolution is not of particular concern, as we are more<BR>> > interested in the cells as whole entities rather than e.g. the details<BR>> > of virus assembly in each cell. However, we would like comparable<BR>> > resolution in all three dimensions, as there is no real "plane" we are<BR>> > interested in and the cells can go in any direction.<BR>> ><BR>> > 5. The hydrogel can be fairly optically dense, and we experience<BR>> > prohibitive distortion at anything deeper than ~100 µm with widefield<BR>> > single photon imaging.<BR>> ><BR>
> > 6. Of course, temperature control and media perfusion will be required.<BR>> > Given the nature of the pathogen, we will need to be able to fully<BR>> > decontaminate the sample chamber and anything that has come into contact<BR>> > with the media. We've been looking at Ernst Stelzer's TC-LSFM to get<BR>> > ideas for the tissue culture chamber. We have access to a fabrication<BR>> > service at our institute that can help us design and build the chamber.<BR>> ><BR>> ><BR>> > Just to give you a general idea, we have about $100K available to us to<BR>> > purchase parts for this, not including the fabrication fee for the<BR>> > culture chamber for which we have other money set aside.<BR>> ><BR>
> > Other than general ideas of what direction we should be heading in, or<BR>> > links to papers for setups we could emulate, I also have some more<BR>> > specific questions:<BR>> ><BR>> > A. Would two-photon imaging be an inevitable requirement given the<BR>> > density of the sample?<BR>> ><BR>> > B. If the sort of speed of acquisition we are looking for is beyond<BR>> > reach for any system we could feasibly build ourselves (especially given<BR>> > the need for exposing in multiple fluorescence channels at each<BR>> > position), what should we sacrifice in our experimental plan to make it<BR>> > more within reach?<BR>> ><BR>> > C. Does anyone know of any labs in the Northeast US/Quebec that have<BR>
> > implemented a similar setup that we could go see in person? We are based<BR>> > in Vermont.<BR>> ><BR>> ><BR>> > I appreciate the work everyone on this list puts in for newcomers, so<BR>> > thank you!<BR>> ><BR>> > Best regards,<BR>> > Mel<BR>> ><BR>> ><BR>> > --<BR>> > Menelaos Symeonides<BR>> > Post-Doctoral Associate, Thali Lab<BR>> > Department of Microbiology and Molecular Genetics<BR>> > University of Vermont<BR>> > 318 Stafford Hall<BR>> > 95 Carrigan Dr<BR>> > Burlington, VT 05405<BR>> > <A href="mailto:msymeoni@uvm.edu"><FONT color=#0000ff>msymeoni@uvm.edu</FONT></A><BR>> > Phone: 802-656-1161<BR>> ><BR>
> > _______________________________________________<BR>> > OpenSPIM mailing list<BR>> > <A href="mailto:OpenSPIM@openspim.org"><FONT color=#0000ff>OpenSPIM@openspim.org</FONT></A><BR>> > <A href="http://openspim.org/mailman/listinfo/openspim"><FONT color=#0000ff>http://openspim.org/mailman/listinfo/openspim</FONT></A><BR>> > _______________________________________________<BR>> > OpenSPIM mailing list<BR>> > <A href="mailto:OpenSPIM@openspim.org"><FONT color=#0000ff>OpenSPIM@openspim.org</FONT></A><BR>> > <A href="http://openspim.org/mailman/listinfo/openspim"><FONT color=#0000ff>http://openspim.org/mailman/listinfo/openspim</FONT></A><BR>><BR>> _______________________________________________<BR>> OpenSPIM mailing list<BR>> <A href="mailto:OpenSPIM@openspim.org">
<FONT color=#0000ff>OpenSPIM@openspim.org</FONT></A><BR>> <A href="http://openspim.org/mailman/listinfo/openspim"><FONT color=#0000ff>http://openspim.org/mailman/listinfo/openspim</FONT></A></FONT><BR> </DIV>
<DIV><FONT size=2 face="Segoe UI">
<DIV> </DIV>
<DIV></DIV>
<DIV><BR> </DIV></FONT>
<STYLE> #sig a:link{color: #00579b;} #sig a:visited{color: #00579b;} #sig a:hover{color: #00579b;} </STYLE>
<DIV id=sig style="FONT-SIZE: 10px; BORDER-TOP: #999 1px dotted; FONT-FAMILY: 'Lucida Grande', Verdana, Arial, Sans-Serif; BORDER-BOTTOM: #999 1px dotted; COLOR: #555; PADDING-BOTTOM: 6px; PADDING-TOP: 6px; PADDING-LEFT: 6px; MARGIN: 6px 0px; LINE-HEIGHT: 14px; PADDING-RIGHT: 6px">
<TABLE width=500 border=0>
<TBODY>
<TR>
<TH height=88 width=125 scope=col><IMG alt=ASI src="http://www.ASIimaging.com/images/ASIlogo.jpg" width=100 align=left height=72></TH>
<TH vAlign=top width=365 scope=col align=left>
<P><FONT size=2>Jon Daniels</FONT><FONT size=2><BR><EM><A href="http://www.ASIimaging.com">Applied Scientific Instrumentation</A></EM><BR></FONT><FONT size=1><EM>29391 West Enid Rd, Eugene, OR 97402</EM><BR><EM>Phone: (541) 461-8181</EM> x118<BR><EM>US & Canada: (800) 706-2284</EM><BR><EM>Fax: (541) 461-4018</EM></FONT></P></TH></TR></TBODY></TABLE></DIV></DIV></FONT></BODY></HTML>