Different results in Turnover

[littlemandy]

I get qualitatively different results if I use the separate turnover() function than if I use the Metacommunity() function. The results for coherence and boundary clumping are the same. The only difference is with the turnover function, specifically the simulated mean. Which function is providing the correct answer? Thank you for your help! Example below:

Metacommunity(veg, scores = 1, method = "r1", sims = 1000,

• order = TRUE, allowEmpty = TRUE, binary = TRUE, verbose = FALSE)

$Turnover is different using the separate function: it is significantly high turnover z pval "1,316,771" "-9.12140545737929" "7.41563066350554e-20" simulatedMean simulatedVariance method "298,240.212" "111663.79926418" "r1"

Turnover(veg, method = "r1", sims = 1000, scores = 1, order = TRUE,

• allowEmpty = TRUE, binary = TRUE, verbose = FALSE) turnover z pval simulatedMean simulatedVariance method 1 1,254,166 11.5047 1.249268e-30 1,700,020 38754.11 r1 This is significantly low turnover.

[littlemandy]

Another piece of data: I tested both functions with an idealized data set. The function embedded withing Metacommunity() seems to be the incorrect version. Results:

metaHighT<-Metacommunity(High, scores = 1, method = "r1", sims = 1000,

• order = TRUE, allowEmpty = TRUE, binary = TRUE, verbose = FALSE) There were 50 or more warnings (use warnings() to see the first 50) metaHighT $Comm a b c d e f g h i j k l m n o p q r s t 24 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 22 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 18 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 17 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 16 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 15 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 14 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 13 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 12 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 11 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 10 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 9 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 8 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 6 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

$Coherence embAbs z pval "0" "11.2343731839505" "2.76448874699434e-29" simulatedMean simulatedVariance method "175.086" "15.5848481382236" "r1"

$Turnover turnover z pval "3500" "1.32009958093428" "0.186801772678876" simulatedMean simulatedVariance method "4563.703" "805.774818326342" "r1"

$Boundary index P df 1 0.5142857 0 17

Turnover(High, method = "r1", sims = 1000, scores = 2, order = TRUE,

• allowEmpty = TRUE, binary = TRUE, verbose = FALSE) turnover z pval simulatedMean simulatedVariance method 1 6775 -8.510826 1.726984e-17 5157.794 190.0175 r1

[taddallas]

Thank you for reporting this. I think there was a bug in the Metacommunity function that I may have fixed.

tests <- TestMatrices
mTurn <- plyr::ldply(tests, function(x){Metacommunity(x)$Turnover})
tTurn <- plyr::ldply(tests, Turnover)

This is the table from the Metacommunity results

.id	turnover	z	pval	simulatedMean	simulatedVariance	method
lei1a	1031	0.305496732981808	0.759987857825579	1103.633	237.753770035653	r1
lei1b	116	0.422251289577714	0.672841608274015	130.158	33.5297969466456	r1
lei2a	0	2.4986461558013	0.012466872222379	68.482	27.4076422709954	r1
lei2b	191	-2.3841118266189	0.0171203990500571	114.407	32.1264292827332	r1
lei3c	775	-2.91552413684608	0.00355091639218749	514.982	89.1839641160644	r1
pres3c	0	3.08811212870848	0.00201432452100491	624.234	202.140976098905	r1
pres4a	1960	0.835555902710837	0.403404785332221	2351.89	469.017092367573	r1

This is the table of the Turnover results. They still are a little off, but that may just be due to the randomization procedure.

.id	turnover	z	pval	simulatedMean	simulatedVariance	method
lei1a	1031	1.0365220	0.2999587	1143.742	108.76952	r1
lei1b	116	0.0668101	0.9467329	117.382	20.68550	r1
lei2a	0	3.7354949	0.0001873	84.777	22.69498	r1
lei2b	191	-1.1475440	0.2511569	157.631	29.07862	r1
lei3c	775	-9.0598635	0.0000000	500.560	30.29185	r1
pres3c	0	9.2323839	0.0000000	1484.674	160.81155	r1
pres4a	1960	-4.8881089	0.0000010	1560.889	81.64937	r1

I've included a new argument that allows the seeding of the null matrices, so results should be repeatable, but I am still a bit unsure why the small differences between the two functions. I'll look more into it. For now, you can play around with the new code by installing the package from Github (devtools::install_github('taddallas/metacom')). Let me know if you find any more bugs. 🐛

Cheers, Tad

[littlemandy]

Thank you very much for being so speedy!

[taddallas]

No worries. Let me know if you find any other issues. I'll close this issue for now, but will keep looking into the slight difference between the two functions.

[MaruLacoretz]

Hi, I am working with metacommunities and I am using the metacom package, although, I have had some troubles with it. I had inconsistent outcomes when i run “Metacommunity” and “Turnover” function with the same matrix (12 sites, 167 species). I hope you can guide me to the solution. These was what I have obtained:

ems.test12x167 <- Metacommunity(tabla.prueba7, method='r1', sims=100, scores=1) ems.test12x167 $Coherence embAbs z "479" "9.69942151349105" pval simulatedMean "3.03212861126114e-22" "821.88" simulatedVariance method "35.3505618374337" "r1"

$Turnover turnover z "32362" "-2.95682517632121" pval simulatedMean "0.00310824295169412" "23859.81" simulatedVariance method "2875.44561920233" "r1"

$Boundary index P df 1 7.54902 1.173583e-07 164

VS

turnover.12x167 <- Turnover(tabla.prueba7, method='r1', sims=100, scores=1) turnover.12x167 turnover z pval simulatedMean 1 32362 7.459905 8.658489e-14 54397.91 simulatedVariance method 1 2953.913 r1

The problem is that I have opposite outputs and I can not decide if i have nestedness or turnover (there is an important difference in the simulatedMean). I tried with more sims and with order FALSE but I obtained the same results. A slight different outcome is in the package example too. The simulatedMean is always higher with “turnover” function and here, test is significative at first, but then, no significative. intmat=TestMatrices[[4]] intmat

ems.test <- Metacommunity(intmat, method='r1', sims=100, scores=1) ems.test

$Coherence embAbs z "0" "4.64782364967428" pval simulatedMean "3.35455507360718e-06" "17.58" simulatedVariance method "3.78241545400114" "r1"

$Turnover turnover z "191" "-2.3854439043864" pval simulatedMean "0.0170585237238189" "113.71" simulatedVariance method "32.4006780699716" "r1"

turnover.intmat <- Turnover(intmat, method='r1', sims=100, scores=1, binary=TRUE) turnover.intmat turnover z pval simulatedMean 1 191 -1.165018 0.2440116 160.23 simulatedVariance method 1 26.4116 r1

I am sorry for the inconvenience, but I do not know If I am doing something wrong. If I have to chose a function (Metacommunity or Turnover), which one do you recommend that I follow? Thank you very much! Mariela

[taddallas]

Mariela,

Thank you for pointing this out! The Metacommunity function was incorrect, in that it "filled in" all the embedded absences in the null matrices, which it shouldn't be doing. I think was the result of a poor fix from this same bug. The Turnover function is correct. The Github repo is updated, and I'll record the bug and submit an update to CRAN soon.

intMat <- TestMatrices[[4]]

Metacommunity(intMat, method='tswap', sims=1000)
Turnover(intMat, method='tswap', sims=1000)

Metacommunity output

turnover	z	pval	simulatedMean	simulatedVariance	method
191	-4.286965	1.811307e-05	145.185	10.68705	tswap

Turnover output

turnover	z	pval	simulatedMean	simulatedVariance	method
191	-4.286965	1.81e-05	145.185	10.68705	tswap

[MaruLacoretz]

Thank you so much for your answer! I've just run it successfully!