Related to tuning of Repeat time

[subhasiny]

I would like to adjust the repeat time for analyzing the reads. In 'deep simulator 1.0' paper as attached in the pdf, there is a parameter in page number 7 called 'a' for adjusting. But I couldn't find the same in Deep simulator 1.5. May I know is there any particular way to adjust the repeat time? I tried to add it at code level in kmer_simulator.py. But I couldn't observe any changes.

Please let us know whether the privilege of altering repeat time is available in 1.5 version and if so, how to perform changes?

[liyu95]

Hi, thank you very much for pointing out that!

Yes, in fact, the 'a' parameter was preserved. You can check the pore_model/src/kmer_simulator.py file.
But we did not include that in the deepsimulator.sh file. You can add it to the file manually. Please let me know if you have any more questions. We will also correct this in our next version.

[subhasiny]

Thank you so much for the reply.

I have varied the value 'a' in kmersimulator.py but my inference was I couldn't get variation with repeat time. Let me give an example so that you can validate my understanding.

For example, I am having a sequence of length 1014. Repetition of each of 1014 nucleotides is determined using 'a' and because of that repetition, I could get an output signal with a length of 8574 (8 to 9 times increase) (as shown in the signal file). Can I vary the length to 10,000 by adjusting repeat_alpha or repeat_more?

Regards
Subhasiny S

[liyu95]

It's possible to do that. You can change the parameter "repeat_more". By varying that, you can change the repeat times. But the distribution may not fit the real data well.

Sincerely,
Yu

[subhasiny]

I have a sequence of length 161. The real nanopore signal length is 3000. But with deepsimulator, i can generate signal of length 1387. After adjusting "repeat_more" parameter, I didn't get any modification in signal length.

When I warped the signal with DWT, both (nanopore and deepsimulator's signal) looks similar.

Is there any parametric changes that I can do to get similar signal with deepsimulator itself?

Thanks
Subhasiny S

[subhasiny]

In function called 'rep_rvs', 'more' parameter is used. However, the action to be taken if 'more' is not equal to one was not written. Can I know if that links to non-alteration of signal repeat length on varying those parameter?

I have attached the permalink for your reference.

DeepSimulator/pore_model/src/poremodel_util.py

Line 68 in d5113cc

def rep_rvs(size,a, more, seed=0):