"MaxEstLhood" is always "-inf"

[Hongpu Chen]

Dear FSC users, 

I am using the fastsimcoal2 (version: 2.6.0.3) to find out which population diverged first, and then which one diverged next.

Down below is one model I wrote, defining the D_E pop diverged first.

tpl:

//Number of population samples (demes)

3

//Population effective sizes (number of genes)

D_E

D_ONE

D_TWO

//Sample sizes

128

112

188

//Growth rates : negative growth implies population expansion

0

0

0

//Number of migration matrices : 0 implies no migration between demes

0

//historical event: time, source, sink, migrants, new size, new growth rate, migr. matrix

2 historical event

TDIV_W1W2 2 1 1 RESIZE_W1W2 0 0

TDIV_E_W1W2 0 1 1 RESIZE_ALL 0 0

//Number of independent loci [chromosome]

1 0

//Per chromosome: Number of linkage blocks

1

//per Block: data type, num loci, rec. rate and mut rate + optional parameters

FREQ 1 0 2.5e−9 OUTEXP

est:

// Priors and rules file

// *********************

[PARAMETERS]

//#isInt? #name #dist. #min #max

//all N are in number of haploid individuals

1 D_E unif 1e4 5e6 output

1 D_ONE unif 1e4 5e6 output

1 D_TWO unif 1e4 5e6 output

1 N_TWO unif 1e4 5e6 output

1 N_ALL unif 1e4 5e6 output

1 TDIV_W1W2 unif 1e5 1e7 output

1 TDIV_E_W1W2 unif 1e6 1e8 ouput

[RULES]

TDIV_E_W1W2>TDIV_W1W2

[COMPLEX PARAMETERS]

0 RESIZE_W1W2 = N_TWO/D_ONE output

0 RESIZE_ALL = N_ALL/N_TWO output

The effectiv population sizes, and the approximately divergence time (a period) I set are according to the results of PSMC. I have already run it several times, what confused me is that each MaxEstLhood of each run is "-inf". I check the "brent_lhoods", all of which were "-inf", and the values of effective population of corresponding pop in "maxL.par" were pairedly the same.

I will really appreciate it, if you could give some advice.

Here are some screenshots and text:

(1) bestlhoods

(2) maxL.par

//Number of population samples (demes)

3

//Population effective sizes (number of genes)

10201

4942723

4942723

//Sample sizes

128

112

188

//Growth rates : negative growth implies population expansion

0

0

0

//Number of migration matrices : 0 implies no migration between demes

0

//historical event: time, source, sink, migrants, new size, new growth rate, migr. matrix

2 historical event

589005437 2 1 1 2.063842137218695e-03 0 0

622412684 0 1 1 1.000000000000000e+00 0 0

//Number of independent loci [chromosome]

1 0

//Per chromosome: Number of linkage blocks

1

//per Block: data type, num loci, rec. rate and mut rate + opt    ional parameters

FREQ 1 0 2.5e−9 OUTEXP

[Hongpu Chen]

I would like to add some infomation here:

the total number of SNP I used in this analysis, is 2755053, but the monomorphic sites is 2744991. Is this because there are too many 0 in my input multi-SFS?

[Aleix Palahí]

Hey!

I am having the same issue with my models, was wondering if you figured out a solution to it? What was causing the fault?

El dia dilluns, 20 de setembre de 2021 a les 3:35:19 UTC+2, Hongpu Chen va escriure:

[Ecstasia Tisiphoni]

Had the same issue, and for me it was the '-' in the mutation rate. It was a dash instead of a hyphen (not very visible) and caused all likelihoods to be '-Inf'.

Hope that helps (even though late).