search

popgenmethods / smcpp

SMC++ infers population history from whole-genome sequence data.
GNU General Public License v3.0
152 stars 35 forks source link

smcpp results weired #208

Open Wennie-s opened 2 years ago

Wennie-s commented 2 years ago

Hi, I am running smcpp, however, the result is very weired. I have 51 individuals in one pop. and the result is image . However, In the wild, the real Ne of this species is very large. Based on the transcriptome data, we used the stairway plot to calculate the Ne, and showed that this species would experienced a bottleneck effect in 1 million years ago, but in 0.6 million years, Ne will rebound. I don't know why the result calculated by smcpp is so strange. Can you help me? the command is: (1)smc++ vcf2smc Chr10_1.MAFfilter1.vcf.gz /wtmp/user003/Reseq/83samples/Ne/smc++/Cudu/smc_gz/Chr10_1.smc.gz Chr10_1 Cudu:DQ-1-2,DQ-1-5,DQ-1-6,DQ-2-1,DQ-2-10,DQ-2-6,DQ-3-1-merged,DQ-3-5,DQ-3-8,DQ-4-1,DQ-4-2-merged,DQ-4-3,DSX-1-3,DSX-1-4,DSX-1-6,DSX-1-8,KG-1,KG-3,KG-4,KG-6,KS-1,KS-4,KS-7,JC-2-3,JC-2-8,LJ-3-2,M64-1,HTX-1,HTX-8,HZ-1,HZ-5,DSC-1,DSC-4,BD-1,M64-5B,M65-1,M66-1,ML-1-10,ML-1-4,ML-8,MZ-1-4,MZ-1-8,MZ-2-2,MZ-3-1,MZ-3-4,PP-1-1,PP-1-6,WS-1-5,WS-1-7,XC-1-3,XLHS-8 (2)smc++ estimate -o Cudu_out 7e-9 /wtmp/user003/Reseq/83samples/Ne/smc++/Cudu/smc_gz/Chr*.smc.gz (3)smc++ plot -g 50 plot.pdf Cudu_out/model.final.json

chen1238 commented 2 years ago

Maybe you have too many samples. i just use 10 samples

elizeng commented 2 years ago

I had the same issue, and after increasing the knot value from the default of 8 to 24 I got more resolution for my plot. Although the general trend remained the same.

Devashish13 commented 2 years ago

1) in the estimation command you have not provided timepoints option, so smc++ automatically tries to identify timepoints where coaelescent events may occur which is weird most of the times. 2) in the plot command you have used generation time as 50 years which is quite high i doubt going back in time it will be that high. 3) Also how far you can go back in time that is restricted by the demography of that species, if there have been recent bottlenecks the coalescent event occurs recently and thus the effective population size is small.