[R] 3 dimensional MDS plots

[Sarah Goslee]

Hi Calla,

On Thu, Mar 10, 2011 at 3:38 PM, Calla Carbone <6cc4 at queensu.ca> wrote:
> Hi,
>
> I am trying to create 3 mds plots: one with axis 1 vs axis 2, one with axis 2 vs axis 3, and one with axis 1 vs axis 3. When inputting my code, I only end up with one diagram and when inputting
> mdsg.mds$dims, the program returns 2 for 2 dimensions. How can I create the other two plots?
>

Assuming you are using metaMDS from the vegan package (it's important
to specify what package
you are using if your function is not in base R), the default option
is to use 2 dimensions. Since you
don't specify k, you only get 2 dimensions.

For NMDS, unlike most other ordination methods, you  need to choose
the number of dimensions
*first*.

mdsg.mds<- metaMDS(mdsg, k=3)

Sarah

> Any help would be greatly appreciated,
> Calla Carbone
>
>
> The table I use is number of each taxa on each surface and is attached.
>
>
>
>
> Below is a copy of my work:
>
>>mdsg=read.csv(file.choose(), header=TRUE, row.names=1, sep=",")
>> mdsg.mds<- metaMDS(sp.rich)
> Error in metaMDS(sp.rich) : object 'sp.rich' not found
>> mdsg.mds<- metaMDS(mdsg)
> Square root transformation
> Wisconsin double standardization
> Run 0 stress 10.62813
> Run 1 stress 14.73774
> Run 2 stress 11.19939
> Run 3 stress 10.61520
> ... New best solution
> ... procrustes: rmse 0.005646588  max resid 0.009332332
> Run 4 stress 10.61520
> ... New best solution
> ... procrustes: rmse 0.0001009763  max resid 0.0001754211
> *** Solution reached
>
>> mdsg.mds.ALT<- metaMDS(mdsg, distance="euclidean", k=3, trymax=50, autotransform=FALSE)
> Run 0 stress 3.020817
> Run 1 stress 0.00673153
> ... New best solution
> ... procrustes: rmse 0.09349157  max resid 0.2443702
> Run 2 stress 0.008861757
> ... procrustes: rmse 0.08275172  max resid 0.1932439
> Run 3 stress 0.006908548
> ... procrustes: rmse 0.0447681  max resid 0.1093436
> Run 4 stress 0.007621072
> ... procrustes: rmse 0.05445531  max resid 0.1437309
> Run 5 stress 0.005675308
> ... New best solution
> ... procrustes: rmse 0.1021018  max resid 0.2694876
> Run 6 stress 0.008473566
> ... procrustes: rmse 0.1411485  max resid 0.3583789
> Run 7 stress 0.008235768
> ... procrustes: rmse 0.05178896  max resid 0.1487678
> Run 8 stress 0.009192125
> ... procrustes: rmse 0.1135218  max resid 0.3126878
> Run 9 stress 0.00984179
> ... procrustes: rmse 0.1117180  max resid 0.3110542
> Run 10 stress 0.00942402
> ... procrustes: rmse 0.08053243  max resid 0.2261970
> Run 11 stress 0.008956116
> ... procrustes: rmse 0.06461741  max resid 0.1729917
> Run 12 stress 0.00871519
> ... procrustes: rmse 0.08382491  max resid 0.2297745
> Run 13 stress 0.006883284
> ... procrustes: rmse 0.05259355  max resid 0.1339228
> Run 14 stress 0.007745699
> ... procrustes: rmse 0.1032920  max resid 0.2410554
> Run 15 stress 0.009062592
> ... procrustes: rmse 0.07324642  max resid 0.2127011
> Run 16 stress 0.006311344
> ... procrustes: rmse 0.1242246  max resid 0.3201084
> Run 17 stress 0.00679012
> ... procrustes: rmse 0.05894517  max resid 0.1699838
> Run 18 stress 0.00892886
> ... procrustes: rmse 0.0886889  max resid 0.2505874
> Run 19 stress 0.007831356
> ... procrustes: rmse 0.04815574  max resid 0.1443481
> Run 20 stress 0.008307666
> ... procrustes: rmse 0.08109922  max resid 0.2286835
> Run 21 stress 0.009738125
> ... procrustes: rmse 0.04494221  max resid 0.0853707
> Run 22 stress 0.006648893
> ... procrustes: rmse 0.0714659  max resid 0.180198
> Run 23 stress 0.008319666
> ... procrustes: rmse 0.1326991  max resid 0.3451491
> Run 24 stress 0.007369039
> ... procrustes: rmse 0.1406218  max resid 0.3560638
> Run 25 stress 0.006693668
> ... procrustes: rmse 0.06047182  max resid 0.1445884
> Run 26 stress 0.007021456
> ... procrustes: rmse 0.06223025  max resid 0.1552082
> Run 27 stress 0.008450277
> ... procrustes: rmse 0.06290812  max resid 0.1474387
> Run 28 stress 0.007949867
> ... procrustes: rmse 0.06185912  max resid 0.1779437
> Run 29 stress 0.007607591
> ... procrustes: rmse 0.0791691  max resid 0.2204299
> Run 30 stress 0.007450255
> ... procrustes: rmse 0.1027691  max resid 0.2798933
> Run 31 stress 0.006476469
> ... procrustes: rmse 0.02180973  max resid 0.04587615
> Run 32 stress 0.008423953
> ... procrustes: rmse 0.05009894  max resid 0.1326238
> Run 33 stress 0.009818617
> ... procrustes: rmse 0.1293366  max resid 0.3352202
> Run 34 stress 0.009782313
> ... procrustes: rmse 0.02056321  max resid 0.04534656
> Run 35 stress 0.007691246
> ... procrustes: rmse 0.06954554  max resid 0.1907627
> Run 36 stress 0.009071507
> ... procrustes: rmse 0.03511019  max resid 0.08272775
> Run 37 stress 0.00867032
> ... procrustes: rmse 0.03915756  max resid 0.08640346
> Run 38 stress 0.008109802
> ... procrustes: rmse 0.04896411  max resid 0.09568495
> Run 39 stress 0.0065526
> ... procrustes: rmse 0.07546078  max resid 0.1902766
> Run 40 stress 0.007617851
> ... procrustes: rmse 0.02149247  max resid 0.0493953
> Run 41 stress 0.007433552
> ... procrustes: rmse 0.03954261  max resid 0.07601604
> Run 42 stress 0.008697543
> ... procrustes: rmse 0.1110775  max resid 0.2855709
> Run 43 stress 0.006243913
> ... procrustes: rmse 0.0806811  max resid 0.2283781
> Run 44 stress 0.008474417
> ... procrustes: rmse 0.1519882  max resid 0.3552259
> Run 45 stress 0.0083211
> ... procrustes: rmse 0.0727077  max resid 0.1767035
> Run 46 stress 0.008051925
> ... procrustes: rmse 0.06954848  max resid 0.172751
> Run 47 stress 0.009271762
> ... procrustes: rmse 0.09332512  max resid 0.2182002
> Run 48 stress 0.009197149
> ... procrustes: rmse 0.104169  max resid 0.2856122
> Run 49 stress 0.00656145
> ... procrustes: rmse 0.04243012  max resid 0.1266379
> Run 50 stress 0.006636944
> ... procrustes: rmse 0.09853805  max resid 0.2584870
>> names(mdsg.mds)
> [1] "points"    "dims"      "stress"    "data"      "distance"  "converged"
> [7] "tries"     "species"   "call"
>> mdsg.mds$points
>             [,1]        [,2]
> Disc  -0.75970137  0.82615812
> Johns  1.09698866 -0.42726075
> LBW   -0.69603453 -0.21063881
> Spbay -0.11433730 -0.03734453
> BC     0.80952315  0.09556815
> D      0.50518560 -0.24053747
> E      0.22159986  0.04797842
> G     -0.40841770 -0.27384498
> LMP    0.23686884  1.23858885
> PC    -0.87625155 -0.64825212
> SH    -0.01542367 -0.37041488
> attr(,"centre")
> [1] TRUE
> attr(,"pc")
> [1] TRUE
> attr(,"halfchange")
> [1] TRUE
>> mdsg.mds$dims
> [1] 2
>> mdsg.mds$stress
> [1] 10.61520
>> mdsg.mds$data
> [1] "wisconsin(sqrt(mdsg))"
>> mdsg.mds$distance
> [1] "bray"
>> mdsg.mds$converged
> [1] TRUE
>> mdsg.mds$tries
> [1] 4
>> mdsg.mds$species
>                        [,1]        [,2]
> Charnia          -0.64803962 -0.25570377
> Charniodiscus    -0.29435409  0.05427335
> Ivesheadia       -0.31324619 -0.33485080
> Thectardis       -0.90038964 -0.63829244
> Tree.Frond       -0.52192114 -0.15500003
> Fractofusus       0.77371091 -0.11631453
> Bradgatia        -0.04960810 -0.21083582
> Pectinifrons      0.04014456 -0.32244681
> Aspidella        -0.45120914  0.39904260
> Hapsidophylla     0.50383940 -0.17459408
> Heimalora        -0.10441330  0.62593841
> Ostrich.Feather   0.28843372  1.46582729
> Feather.Duster    0.01804479 -0.02227495
> Tulip.Frond       0.66742592  0.73062070
> Trepassia         0.28843372  1.46582729
> Beothukis         0.73931198 -0.08363615
> Primocandelabrum -0.92508363  0.97772972
> Hadryniscala     -0.92508363  0.97772972
> Blackbrookia      1.33579627 -0.50564841
> Frondophyllas     0.28843372  1.46582729
> Parviscopa       -0.92508363  0.97772972
> attr(,"shrinkage")
> [1] 0.6744097 0.7139844
> attr(,"centre")
> [1]  1.788934e-18 -3.794708e-19
> attr(,"old.wa")
> [1] FALSE
>> mdsg.mds$call
> metaMDS(comm = mdsg)
>> mds1.mds
>
> Call:
> metaMDS(comm = mds1)
>
> Nonmetric Multidimensional Scaling using isoMDS (MASS package)
>
> Data:     wisconsin(sqrt(mds1))
> Distance: bray
>
> Dimensions: 2
> Stress:     9.804838
> Two convergent solutions found after 3 tries
> Scaling: centring, PC rotation, halfchange scaling
> Species: expanded scores based on ‘wisconsin(sqrt(mds1))’
>
>> variableScores<- mds1.mds$species
>> sampleScores<- mds1.mds$points
>> plot(mds1.mds)
>> plot(mds1.mds, type="n")
>> text(mds1.mds, display=c("sites"), choices=c(1,2), col="blue", cex=0.7)
>> points(mds1.mds, display=c("species"), choices=c(1,2), pch=3, col="red")
>
>

-- 
Sarah Goslee
http://www.functionaldiversity.org