Re: [Nemo2-user] Nemo installation difficulties with Cygwin
Individual-based forward-time genetics simulation software
Status: Beta
Brought to you by:
freg
From: frederic g. <gui...@gm...> - 2016-02-24 16:19:28
|
Hi David, Binary files are necessary to source quantitative traits, there is no reader for .quanti files yet. best Fred On Wed, 2016-02-24 at 11:12 -0500, David Remington wrote: > Hi Fred, > > > Does the quantitative trait component in Nemo allow specifying a > source population? If so, does it have to be a binary file, or can it > read the .quanti files? > > > Thanks, > > Dave > > > > > David L. Remington > Associate Professor > Department of Biology > University of North Carolina at Greensboro > P.O. Box 26170 > Greensboro, NC 27402-6170 > > lab and office: 226 Eberhart Bldg. > tel: (336) 334-4967 > e-mail: dlr...@un... > web site: http://biology.uncg.edu/people/David_Remington/ > > On Thu, Dec 17, 2015 at 11:21 AM, David Remington <dlr...@un...> > wrote: > Thanks! Changing the fitness model fixed the problem we were > having, and I can now simulate 200 loci. We'll also be > incorporating a bunch of two-population models with different > fitness optima, and will need to play around with the effects > of different parameter values for fitness optima, selection > strength, and environmental variance anyway, so I think the > Gaussian model will work fine. I do think the Gaussian model > is more biologically realistic anyway when the population is > some distance from the optimum, since fitness will approach > zero asymptotically. > > > Happy holidays! > > > Dave > > > > > > > David L. Remington > Associate Professor > Department of Biology > University of North Carolina at Greensboro > P.O. Box 26170 > Greensboro, NC 27402-6170 > > lab and office: 226 Eberhart Bldg. > tel: (336) 334-4967 > e-mail: dlr...@un... > web site: http://biology.uncg.edu/faculty/David_Remington/ > > On Thu, Dec 17, 2015 at 7:45 AM, frederic guillaume > <gui...@gm...> wrote: > The quadratic model is an approximation of the > Gaussian when the > population sits close to its optimum, which is likely > the case in your > scenario with a single population and a small > mutational variance > > you might have seen decimal values for recombination > *rates* per > chromosome, I'll add a warning in a next version to > avoid the confusion > > thanks for the feedback and helping make nemo better! > > best, and happy holidays > Fred > > On Tue, 2015-12-15 at 16:25 -0500, David Remington > wrote: > > Thanks for the suggestions! It seems to me that the > Gaussian > > selection model would change the scaling of the > fitness costs relative > > to the quadratic model, but I can certainly live > with that, and the > > Gaussian model does seem more biologically > realistic. > > > > > > I think I had initially understood that the map > position values needed > > to be integers based on the manual, but then I > noticed one of the > > example files had decimal values. I guess that was > from an earlier > > version, though. I'll make those changes!! > > > > > > Thanks, > > > > Dave > > > > > > > > > > David L. Remington > > Associate Professor > > Department of Biology > > University of North Carolina at Greensboro > > P.O. Box 26170 > > Greensboro, NC 27402-6170 > > > > lab and office: 226 Eberhart Bldg. > > tel: (336) 334-4967 > > e-mail: dlr...@un... > > web site: > http://biology.uncg.edu/faculty/David_Remington/ > > > > On Tue, Dec 15, 2015 at 4:00 PM, Fred Guillaume > > <gui...@gm...> wrote: > > Hi > > The problem comes from the quadratic > selection model you chose > > use the "gaussian" option instead, which > will give very close > > results > > I haven't found the bug yet > > > > one other thing, I noticed that you use > decimal values for map > > positions but map position must be integers, > values after the > > decimal points are ignored, > > in your case, multiply all positions by 100 > and change the map > > resolution (scale) to: > > quanti_genetic_map_resolution 0.01 > > which will give you the correct positions > > > > best > > F > > > > > > On 14 déc. 2015, at 17:02, David Remington > wrote: > > > > > Hi Fred, > > > > > > Actually, I seem to be able to use 51 > loci, but when I add > > one more to make it 52, I get the following > error: > > > replicate 1/1 [10:45:19] > 1/100000***ERROR*** > > Individual::Trying to access a trait not > present in the traits > > table (at 32564, size 1) > > > > > > The .ini file is attached. If I were to > remove the very > > last marker (at position 224.08) and reduce > quanti_loci to 51, > > it would again run correctly. > > > > > > Thanks, > > > Dave > > > > > > > > > > > > David L. Remington > > > Associate Professor > > > Department of Biology > > > University of North Carolina at Greensboro > > > P.O. Box 26170 > > > Greensboro, NC 27402-6170 > > > > > > lab and office: 226 Eberhart Bldg. > > > tel: (336) 334-4967 > > > e-mail: dlr...@un... > > > web site: > http://biology.uncg.edu/faculty/David_Remington/ > > > > > > On Mon, Dec 14, 2015 at 3:27 AM, frederic > guillaume > > <gui...@gm...> wrote: > > > Dear David, > > > > > > A genetic map of more than 50 loci is > absolutely possible! > > What troubles > > > are you having? Let me know about your > parameters so I can > > help you with > > > that. > > > > > > Best > > > Fred > > > > > > > > > > > > On Fri, 2015-12-11 at 15:49 -0500, David > Remington wrote: > > > > Dear Fred, > > > > > > > > > > > > I actually have another question > regarding Nemo. (I'll > > just send this > > > > to you directly to avoid issues with you > getting messages > > from the > > > > Nemo-users site.) Is it possible to get > Nemo to accept a > > genetic map > > > > with more than 50 loci? I want to do > some QTL simulations > > with sets > > > > of tightly-clustered loci to simulate > individual genes > > that can > > > > occasionally experience intragenic > recombination, so it > > would be nice > > > > to simulate larger numbers of loci (up > to 200 perhaps). > > However, 50 > > > > loci seems to be the maximum, with more > than that > > generating an error > > > > in the first generation. > > > > > > > > > > > > Thanks, > > > > > > > > Dave > > > > > > > > > > > > > > > > > > > > David L. Remington > > > > Associate Professor > > > > Department of Biology > > > > University of North Carolina at > Greensboro > > > > P.O. Box 26170 > > > > Greensboro, NC 27402-6170 > > > > > > > > lab and office: 226 Eberhart Bldg. > > > > tel: (336) 334-4967 > > > > e-mail: dlr...@un... > > > > web site: > > > http://biology.uncg.edu/faculty/David_Remington/ > > > > > > > > On Fri, Dec 11, 2015 at 9:07 AM, David > Remington > > <dlr...@un...> > > > > wrote: > > > > Thank you for the information, > Fred! > > > > > > > > > > > > Since I posted this request, I > have since used > > VirtualBox to > > > > install a Ubuntu Linux virtual > machine on my > > computer, and > > > > have successfully installed Nemo > and started using > > it in that > > > > way. > > > > > > > > > > > > Dave > > > > > > > > > > > > > > > > > > > > David L. Remington > > > > Associate Professor > > > > Department of Biology > > > > University of North Carolina at > Greensboro > > > > P.O. Box 26170 > > > > Greensboro, NC 27402-6170 > > > > > > > > lab and office: 226 Eberhart > Bldg. > > > > tel: (336) 334-4967 > > > > e-mail: dlr...@un... > > > > web site: > > > http://biology.uncg.edu/faculty/David_Remington/ > > > > > > > > On Fri, Dec 11, 2015 at 2:54 AM, > frederic > > guillaume > > > > <gui...@gm...> > wrote: > > > > Dear David, > > > > > > > > Sorry for the delayed > answer, you message > > hasn't been > > > > properly routed to > > > > my email. > > > > > > > > To compile on Windows > with CygWing, the > > easiest is to > > > > specify WIN=1 on > > > > the command line, like > this: > > > > > > > > $ make WIN=1 > > > > > > > > or > > > > > > > > $ make install WIN=1 > > > > > > > > this should correctly > link nemo with the > > GSL (although > > > > statically) > > > > > > > > if you don't want static > linking, edit > > line 68 of the > > > > makefile > > > > > > > > change: > > > > ifdef WIN > > > > LD_OPTS = -lstdc++ > -static-libgcc > > -static -lgsl > > > > -lgslcblas > > > > else > > > > > > > > to: > > > > ifdef WIN > > > > LD_OPTS = -lstdc++ > -lgsl -lgslcblas > > > > else > > > > > > > > Otherwise, concerning > GSL packages in the > > CygWin > > > > installer, they have > > > > changed names. > > > > > > > > The GSL package is now > in the "science" > > section, which > > > > should install > > > > the libgsl-dev and > libgsl0 packages from > > the "lib" > > > > section. > > > > > > > > I hope this helps! > > > > and sorry again for the > delay > > > > > > > > Best > > > > Fred > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > <DLR_base1Test.ini> > > > > > > > > > > > > |