MCMC Post-hoc Analysis: 48 sequences

Data & Model

Partition	Sequences	Lengths	Alphabet	Substitution Model	Indel Model	Scale Model
1	48-fsa.fasta	110 - 131	RNA	S1 = gtr+Rates.gamma[4]+inv	I1 = rs07	scale1 ~ gamma[0.5,2,0.0]

Scalar variables

Statistic	Median	95% BCI	ACT	ESS	burnin	PSRF-CI80%	PSRF-RCF
prior	-470.3	(-537.4, -406.5)	26.66	53770	279	1	0.9995
prior_A1	-754.2	(-817.1, -696.4)	27.41	52302	279	1	0.9985
likelihood	-4041	(-4082, -4000)	35.26	40651	237	1	0.9982
posterior	-4511	(-4568, -4459)	13.74	104298	228	1	0.9985
Heat.beta	1
Scale[1]	12.36	(8.833, 16.54)	4.88	293739	180	1	1.002
gtr:pi[A]	0.2127	(0.1782, 0.2496)	25.04	57239	572	1	1.001
gtr:pi[C]	0.2822	(0.2475, 0.3183)	13.43	106764	273	1	1.001
gtr:pi[G]	0.2879	(0.2468, 0.3314)	29.92	47899	332	1	1
gtr:pi[U]	0.2152	(0.1844, 0.2469)	28.73	49886	1017	1	1.003
gtr:sym[AC]	0.06996	(0.04235, 0.1)	9.525	150483	353	1	1.003
gtr:sym[AG]	0.205	(0.1577, 0.2554)	18.73	76526	849	1	0.9987
gtr:sym[AU]	0.232	(0.1757, 0.2922)	35.56	40311	886	1	1.001
gtr:sym[CG]	0.09082	(0.06181, 0.1233)	28.22	50790	1020	1	1.005
gtr:sym[CU]	0.3511	(0.2876, 0.4173)	23.65	60595	526	1	0.9992
gtr:sym[GU]	0.04625	(0.02222, 0.0729)	13.07	109695	437	1	0.9984
Rates.gamma:alpha	1.328	(0.7789, 1.873)	12.65	113293	374	0.9999	1.001
inv:p_inv	0.1436	(0.06041, 0.2317)	9.495	150956	205	0.9999	1.002
rs07:mean_length	1.954	(1.608, 2.355)	20.13	71188	389	0.9998	1.001
rs07:log_rate	-3.504	(-3.812, -3.207)	14.41	99487	403	1	1.001
|A1|	201	(186, 215)	169.3	8465	2829	0.9682	0.99
#indels1	101	(90, 111)	26.78	53528	265	0.9286	0.998
|indels1|	189	(163, 213)	49.61	28892	518	0.9734	0.9955
#substs1	1023	(1001, 1045)	46.77	30648	611	0.9707	0.9976
Scale1*|T|	13.75	(11.24, 16.63)	14.13	101470	322	0.9998	1.002
|A|	201	(186, 215)	169.3	8465	2829	0.9682	0.99
#indels	101	(90, 111)	26.78	53528	265	0.9286	0.998
|indels|	189	(163, 213)	49.61	28892	518	0.9734	0.9955
#substs	1023	(1001, 1045)	46.77	30648	611	0.9707	0.9976
|T|	1.116	(0.8416, 1.415)	1.017	1409462	71	1	0.9999

Phylogeny Distribution

Partition support: Summary

Partition support graph: SVG

50% consensus	Newick (+PP)	PDF	SVG
66% consensus	Newick (+PP)	PDF	SVG
80% consensus	Newick (+PP)	PDF	SVG
90% consensus	Newick (+PP)	PDF	SVG
95% consensus	Newick (+PP)	PDF	SVG
99% consensus	Newick (+PP)	PDF	SVG
100% consensus	Newick (+PP)	PDF	SVG
MAP	Newick (+PP)	PDF	SVG
greedy	Newick (+PP)	PDF	SVG

Alignment Distribution

Partition 1

			Diff		Min. %identity	# Sites	Constant	Informative
Initial	FASTA	HTML	Diff		17.7%	214	6 (2.8%)	187 (87.4%)
Best (WPD)	FASTA	HTML		AU	25.8%	206	8 (3.88%)	152 (73.8%)

Mixing

1. Partition uncertainty
2. SRQ plot: partitions
3. SRQ plot: c50
4. Variation in split frequency estimates

burnin (scalar)	ESS (scalar)	ESS (partition)	ASDSF	MSDSF	PSRF-CI80%	PSRF-RCF
2829	8465	10999.325	0.007	0.024	1	1.005

Projection of RF distances for the first 3 chains (Hillis et al 2005) 3D

Variation of split PPs across chains (Beiko et al 2006)

---

Analysis

command line: bali-phy 48-fsa.fasta -S gtr+Rates.gamma+inv

directory: /gpfs/fs0/data/wraycompute/br51/Work/3.1

chain #	burnin	subsample	Iterations (after burnin)	subdirectory
1	19236	1	180764	48-fsa-1
2	19236	1	180764	48-fsa-2
3	19236	1	180764	48-fsa-3
4	19236	1	180764	48-fsa-4
5	19236	1	173128	48-fsa-5
6	19236	1	180764	48-fsa-6
7	19236	1	180084	48-fsa-7
8	19236	1	176297	48-fsa-8

P(data|M) = -4085.760 +- 0.389

Complete sample: 1586668 topologies

95% Bayesian credible interval: 1507308 topologies

Model and priors

Tree (+priors)

topology	~ uniform on tree topologies
branch lengths	~ iid[num_branches[T],gamma[0.5,div[2,num_branches[T]],0.0]]

Substitution model (+priors)

S1

=

gtr+Rates.gamma[4]+inv gtr:sym ~ dirichlet_on[letter_pairs[a],1] gtr:pi ~ dirichlet_on[letters[a],1] Rates.gamma:alpha ~ log_laplace[6,2] inv:p_inv ~ uniform[0,1]

Indel model (+priors)

I1

=

rs07 rs07:log_rate ~ laplace[-4,0.707] rs07:mean_length ~ exponential[10,1]

Scales (+priors)

scale1

~

gamma[0.5,2,0.0]