Closed shangguandong1996 closed 2 years ago
The random scores were generated using the dCNS ranSco
command.
dCNS ranSco
is generating scores for a pair of sequences with 1000bp by default (the -l parameter).
Please forgive me if I misunderstand something
The true mn is 100kb*100kb, why I can use the l
and k
paramter predicted by 1000*1000?
l and k are parameters of the distribution, which could be calculated using shorter sequences.
https://www.ncbi.nlm.nih.gov/BLAST/tutorial/Altschul-1.html Please read this for more details.
Thanks !
Hi, Dr song.
I think the below code in setting parameters manual is using nlsLM to fit the curve produced by
ranSco
But accoring to your equation E = Kmne^(-l*ambdaS) p-value = e^(-E)
so why the y~(1-exp(-1*k*1000000*exp(-1*l*x))) is 1000000(1e6) instead of 1e10(100kb 100kb) ?
Best wishes Guandong Shang