heejungshim / waveqtl Goto Github PK

It seems that pi estimate is not stable when it is estimated using only one WC with very small BF (see Jack's transcript start site QTLs as examples). We should investigate behavior of pi estimate for low-resolution WCs.

i have found that even when I set the -numPerm option larger than 10,000, the largest number of permutations I ever is get is 10000. Is this the maximum number possible?

We could add an option for 2D wavelet decomposition and effect size estimate in 2D image.

WaveQTL sometimes outputs negative logLR. We could handle it as we do in multi-scale method. Permutation procedure should be changed accordingly.