Skip to main content
. 2022 Mar 8;12:789983. doi: 10.3389/fmicb.2021.789983

TABLE 1.

Test results of fitting the HDP-MSN (hierarchical Dirichlet process, multi-site neutral) model of Harris et al. (2017) to the meta-communities consisting of 3-site semen-vaginal samples (CM = Semen Sample, CNA = vaginal sample before intercourse, and CNB = vaginal sample after intercourse) (P-value > 0.05 indicating significant or satisfactory fitting to the MSN)*,**.

ID L O θ m M-value Meta-community
 Local Community
LM N M N P M LL N L N P L
1 –382.709 40.978 0.151 73.153 –396.822 1,566 2,500 0.626 –389.420 1,417 2,500 0.567
2*** −1, 385.486 141.111 0.015 37.188 −1, 370.380 1,073 2,500 0.429 −1, 415.622 1,632 2,500 0.653
3 –708.010 91.888 0.013 35.028 –727.026 1,551 2,500 0.620 –744.482 1,775 2,500 0.710
6 −1, 032.545 100.840 0.020 39.051 −1, 051.866 1,499 2,500 0.600 −1, 070.271 1,803 2,500 0.721
7 −1, 040.569 97.987 0.010 32.693 −1, 069.604 1,594 2,500 0.638 −1, 091.907 1,930 2,500 0.772
8 –753.034 73.124 0.029 59.935 –837.017 2,162 2,500 0.865 –804.300 1,990 2,500 0.796
9 –820.546 106.933 0.012 24.899 –835.507 1,468 2,500 0.587 –855.005 1,776 2,500 0.710
12 –425.321 47.271 0.022 43.152 –460.434 1,819 2,500 0.728 –464.377 1,971 2,500 0.788
14 –717.459 93.516 0.011 33.766 –725.231 1,374 2,500 0.550 –742.193 1,638 2,500 0.655
15 –425.352 40.732 0.040 104.209 –476.039 2,024 2,500 0.810 –448.788 1,722 2,500 0.689
16 –692.768 95.069 0.010 27.169 –730.169 1,788 2,500 0.715 –732.498 1,800 2,500 0.720
17 –874.131 102.123 0.007 18.343 –917.696 1,813 2,500 0.725 –949.243 2,146 2,500 0.858
18 –768.402 97.810 0.011 26.005 –808.089 1,771 2,500 0.708 –819.856 1,986 2,500 0.794
21 −1, 381.209 129.878 0.021 40.338 −1, 392.766 1,381 2,500 0.552 −1, 431.878 1,936 2,500 0.774
22 –690.534 70.839 0.015 40.315 –739.996 1,910 2,500 0.764 –741.027 2,012 2,500 0.805
23 −1, 044.392 49.574 0.041 105.839 −1, 277.726 2,478 2,500 0.991 −1, 099.484 2,078 2,500 0.831
24 –740.746 47.534 0.072 151.599 –880.842 2,428 2,500 0.971 –762.288 1,649 2,500 0.660
25 −1, 262.028 79.649 0.024 76.674 −1, 386.549 2,225 2,500 0.890 −1, 301.471 1,725 2,500 0.690
26 −1, 178.871 96.351 0.016 38.733 −1, 290.164 2,253 2,500 0.901 −1, 256.236 2,161 2,500 0.864
27 –998.888 55.951 0.056 160.489 −1, 127.655 2,303 2,500 0.921 −1, 000.514 1,277 2,500 0.511
28 −1, 059.024 62.918 0.076 207.492 −1, 185.529 2,296 2,,500 0.918 −1, 035.615 820 2,500 0.328
29 −1, 086.036 60.456 0.040 126.032 −1, 339.392 2,490 2,500 0.996 −1, 118.532 1,749 2,500 0.700
30 –981.639 66.499 0.013 35.685 −1, 081.354 2,207 2,500 0.883 −1, 055.134 2,176 2,500 0.870
Mean –889.117 80.393 0.032 66.860 –961.211 1,890 2,500 0.756 –927.397 1,790 2,500 0.716
Passing rate (%) 100% 100%

*N = 2,500 is the number of Gibb samples selected from 25,000 simulated communities (i.e., every tenth iteration of the last 25,000 Gibbs samples), it is chosen to compute the pseudo P-value below for conducting the neutrality test.

L0 is the actual (observed) log-likelihood.

θ is the median of biodiversity numbers computed from 25,000 times of simulations.

m is the migration probability.

M-value is the average medians of the migration rates of local communities in each meta-community (i.e., the average median of the individuals migrated per generation), also computed from 25,000 times of simulations.

LM is the median of the log-likelihoods of the simulated neutral meta-community samples; and NM is the number of simulated neutral meta-community samples with their likelihoods satisfying LM ≤ L0 (where LM and L0 are the simulated and actual likelihood respectively).

PM = NM /N is the pseudo p-value for testing the neutrality at meta-community level; if PM > 0.05, the meta-community is indistinguishable from the prediction of neutral model.

LL is the median of the log-likelihoods of the simulated local community samples, and NL is the number of simulated local community samples with their likelihoods satisfying LL ≤ L0 (where LL and L0 are the simulated and actual likelihood respectively).

PL = NL /N, is the pseudo p-value for testing the neutrality at the local community level; if PL > 0.05, the local community satisfies the neutral model.

**Due to the typo/error in Harris et al. (2017), the PM-values exhibited here are adjusted as (PM = 1−PMS), where PMS is output from their computational program.

Similarly, the PL-values are adjusted as (PL = 1−PLS), where PLS is output from their computational program. ***Figure 2 displayed the fitting of the MSN to #2 sample by plotting the predicted and observed species abundance rank distribution.