Too many steps in routine qromb: Numerical result out of range

[stsmall]

Hi @FerRacimo,
I seem to be running into the above error. Once I get the error, the program continues running and only produces the error output line over and over again. I can provide an input file if it would help.
Here is my command line: 3P-CLR/src/threepclr "${file}" "${file}.output" 20 100 0.0025 0.01353,0.0025,0.4460 NA
thanks,
@stsmall

[FerRacimo]

Hi, You should only use sites that are polymorphic in the outgroup. In other words, if you have any fields where the allele count for the outgroup population is equal to 0 or 100, you should remove those. I suspect that should fix your problem. Best, Fernando

…

On Tue, Feb 16, 2021 at 7:49 PM Scott T Small ***@***.***> wrote: Hi @FerRacimo <https://github.com/FerRacimo>, I seem to be running into the above error. Once I get the error, the program continues running and only produces the error output line over and over again. I can provide an input file if it would help. Here is my command line: 3P-CLR/src/threepclr "${file}" "${file}.output" 20 100 0.0025 0.01353,0.0025,0.4460 NA thanks, @stsmall <https://github.com/stsmall> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLXSATNLGDXCRP3GVH5ZJLS7K44TANCNFSM4XW6XAZQ> .

-- Fernando Racimo, PhD Associate Professor of Population Genetics Globe Institute University of Copenhagen Øster Voldgade 5-7, DK-1350 Copenhagen, Denmark

[stsmall]

Hi @FerRacimo,
Thanks for the reply. When I built the input files I only kept sites that were polymorphic in the outgroup, following what you had specified in the readme. I am not sure why I am getting the error. I have tried a few different things but always end up at the same error. I had some concerns that for snps that are close together (by phypos) share the same genpos value. This because of the linear interpolation from shapeit4, where I input a sparse recombination map for phasing. Could this be the issue? I have attached a set of input files as well as the stdout/err from the same run.
thanks for the support and help!
@stsmall
Xpclr2.o108877.5.gz
2.3pclr.input5.output.gz
2.3pclr.input5.gz

[stsmall]

Hi @FerRacimo,
I think the error was being caused by lines like the following:
X 9998539 44 0 148 0 184 1 18
Where the outgroup is segregating but both test groups are fixed ancestral. Could that be it? So far my test run has not produced the qromb error after removing these sites.
thanks,
@stsmall

[stsmall]

nope. that wasnt it, or it wasnt the cause of all the errors. Seemed to start printing errors on this line:
2 20086080 92.0646 1 148 9 184 2 18

[FerRacimo]

Hi Scott, can you try filtering for sites with >1% frequency in both of the target populations?

…

On Thu, Feb 18, 2021 at 6:15 PM Scott T Small ***@***.***> wrote: nope. that wasnt it, or it wasnt the cause of all the errors. Seemed to start printing errors on this line: 2 20086080 92.0646 1 148 9 184 2 18 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLXSAVE2T2ON5AQ6YZNHDTS7VDMFANCNFSM4XW6XAZQ> .

-- Fernando Racimo, PhD Associate Professor of Population Genetics Globe Institute University of Copenhagen Øster Voldgade 5-7, DK-1350 Copenhagen, Denmark

[stsmall]

Hi @FerRacimo,
Thank you for the continued help.
I only kept sites >1% freq ... still the error:
20
100
0.0025
NA
Windows determined. Beginning statistic computation...
2 8.00002e+07 302.573 8.00002e+07 8.00004e+07 302.573 302.574 50.3488 0.1 49.8751 0.1 49.8717 0.1 1.23936 0.001 31.4585 0.1
2 8.00045e+07 302.595 8.00042e+07 8.00048e+07 302.593 302.596 60.9019 0.1 59.8923 0.1 59.8872 0.1 0.575988 0.0001 31.8225 0.1
2 8.00087e+07 302.62 8.00086e+07 8.00088e+07 302.619 302.621 102.658 0.1 100.39 0.1 100.792 0.1 4.00378 0.0001 63.5417 0.1
2 8.00117e+07 302.638 8.00115e+07 8.00119e+07 302.637 302.639 177.116 0.1 160.697 0.1 160.608 0.1 6.50538 0.0001 95.1303 0.1
2 8.00134e+07 302.648 8.00132e+07 8.00135e+07 302.647 302.649 160.265 0.1 151.901 0.1 151.853 0.1 1.64516 0.0001 96.0489 0.1
2 8.00148e+07 302.657 8.00147e+07 8.0015e+07 302.656 302.658 336.966 0.1 333.115 0.1 333.504 0.1 8.78223e-08 1e-05 222.531 0.1
Too many steps in routine qromb: Numerical result out of range
Too many steps in routine qromb: Numerical result out of range
Too many steps in routine qromb: Numerical result out of range

Here is the input file lines around that error:
2 80014294 302.653 113 148 148 184 15 18
2 80014363 302.654 3 148 7 184 1 18
2 80014387 302.654 80 148 71 184 12 18
2 80014699 302.656 16 148 20 184 6 18
2 80014783 302.656 32 148 35 184 2 18
2 80014789 302.656 4 148 6 184 3 18
2 80014791 302.656 22 148 29 184 8 18
2 80014807 302.657 38 148 47 184 6 18
2 80014852 302.657 46 148 40 184 1 18
2 80014882 302.657 23 148 34 184 8 18
2 80014897 302.657 103 148 134 184 13 16
2 80014900 302.657 42 148 45 184 1 18
2 80014924 302.657 10 148 10 184 1 18

[FerRacimo]

Hmmm I still suspect this has to do with low frequency variants. Try setting a 1% MAF cutoff on the 3rd (outrgroup) population as well. If that doesn't work, try setting a 5% MAF cutoff on all 3 populations.

…

On Thu, Feb 18, 2021 at 8:24 PM Scott T Small ***@***.***> wrote: Hi @FerRacimo <https://github.com/FerRacimo>, Thank you for the continued help. I only kept sites >1% freq ... still the error: 20 100 0.0025 NA Windows determined. Beginning statistic computation... 2 8.00002e+07 302.573 8.00002e+07 8.00004e+07 302.573 302.574 50.3488 0.1 49.8751 0.1 49.8717 0.1 1.23936 0.001 31.4585 0.1 2 8.00045e+07 302.595 8.00042e+07 8.00048e+07 302.593 302.596 60.9019 0.1 59.8923 0.1 59.8872 0.1 0.575988 0.0001 31.8225 0.1 2 8.00087e+07 302.62 8.00086e+07 8.00088e+07 302.619 302.621 102.658 0.1 100.39 0.1 100.792 0.1 4.00378 0.0001 63.5417 0.1 2 8.00117e+07 302.638 8.00115e+07 8.00119e+07 302.637 302.639 177.116 0.1 160.697 0.1 160.608 0.1 6.50538 0.0001 95.1303 0.1 2 8.00134e+07 302.648 8.00132e+07 8.00135e+07 302.647 302.649 160.265 0.1 151.901 0.1 151.853 0.1 1.64516 0.0001 96.0489 0.1 2 8.00148e+07 302.657 8.00147e+07 8.0015e+07 302.656 302.658 336.966 0.1 333.115 0.1 333.504 0.1 8.78223e-08 1e-05 222.531 0.1 Too many steps in routine qromb: Numerical result out of range Too many steps in routine qromb: Numerical result out of range Too many steps in routine qromb: Numerical result out of range Here is the input file lines around that error: 2 80014294 302.653 113 148 148 184 15 18 2 80014363 302.654 3 148 7 184 1 18 2 80014387 302.654 80 148 71 184 12 18 2 80014699 302.656 16 148 20 184 6 18 2 80014783 302.656 32 148 35 184 2 18 2 80014789 302.656 4 148 6 184 3 18 2 80014791 302.656 22 148 29 184 8 18 2 80014807 302.657 38 148 47 184 6 18 2 80014852 302.657 46 148 40 184 1 18 2 80014882 302.657 23 148 34 184 8 18 2 80014897 302.657 103 148 134 184 13 16 2 80014900 302.657 42 148 45 184 1 18 2 80014924 302.657 10 148 10 184 1 18 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLXSATVA5A3NXPMGIGQGDLS7VSPZANCNFSM4XW6XAZQ> .

-- Fernando Racimo, PhD Associate Professor of Population Genetics Globe Institute University of Copenhagen Øster Voldgade 5-7, DK-1350 Copenhagen, Denmark

[stsmall]

Hi @FerRacimo,
Filtering the outgroup to keep only >1% seemed to work. Been running for a few hours without any error.
Thank you very much for all the help while we figured it out.
@stsmall

Yes, that did the trick! Thanks!!!!!