Very low swap rates between certain chains

[MattD]

Hi all,

I'm running a data set with 40 individuals from two populations, data
is mtDNA (1 locus) and STR (10 loci). I'm getting decent swap rates
between most of my chains, but very low swap rates for the lowest-
numbered chains, often right around chain 3. In some runs these drop
as low as 0 swaps for Chain 3.

The heating parameters I've been using are
-hfg -hn80 -ha0.999 -hb0.3

Has anyone else encountered this before? I'm trying some different
heating schemes, but I would appreciate any suggestions on how to
adjust heating parameters to address this.


For example:

CHAIN SWAPPING BETWEEN SUCCESSIVE CHAINS: Geometric Increment term1:
0.9990 term2: 0.3000
-----------------------------------------------------------------------------
Chain beta #Swaps Rate
0 1.0000 889 0.2911
1 0.9918 4 0.0015
2 0.9836 14 0.0058
3 0.9754 2 0.0010
4 0.9671 124 0.0641
5 0.9589 382 0.2119
6 0.9506 543 0.3456
7 0.9423 609 0.3952
8 0.9340 601 0.4020
9 0.9256 784 0.4978
10 0.9173 618 0.4052
11 0.9089 584 0.3797
12 0.9006 578 0.3838
13 0.8922 586 0.3984
14 0.8838 665 0.4451
15 0.8753 584 0.3830
16 0.8669 718 0.4733
17 0.8584 652 0.4370
18 0.8499 628 0.4057
19 0.8415 666 0.4239
20 0.8329 621 0.4213
21 0.8244 648 0.4252
22 0.8159 627 0.4163
23 0.8073 651 0.4139
24 0.7987 542 0.3506
25 0.7901 656 0.4330
26 0.7815 615 0.3991
27 0.7729 652 0.4242
28 0.7642 578 0.3647
29 0.7556 581 0.3832
30 0.7469 648 0.4058


Thank you in advance for any suggestions!

Matt

[jhey]

The curve is too flat. drop the -ha value a little. e.g. try -
ha0.99

jhey