Co-occurence analysis and co-occurrence network.

[Hira_12]

I have the following Table for which i want to do the co-occurence analysis. I want to make a co-occurence network too.
My data structure looks like this

https://docs.google.com/spreadsheets/d/1ML_1fhNdI4kr9XRTcB0o0r9MvrGhc3yZ/edit?usp=sharing&ouid=117110307617159452368&rtpof=true&sd=true

For this analysis in R, I am following this tutorial :

How to create co-occurrence networks with the R packages ‘cooccur’ and ‘visNetwork’ | by Brooke Bradley | Analytics Vidhya | Medium

https://www.researchgate.net/publication/293654442_cooccur_Probabilistic_Species_Co-Occurrence_Analysis_in_R

I don't know how to create a presence/absence matrix for my data. for example like this one:

And finally create a co-occurence network using that matrix.

I have added all the necessary details for which i need help. I am unable to do this analysis on my own as I am new to R. I hope somebody can help me in this.
Regards
Hira

[jms]

Hi Hira,
you can use the ifelse() function to recode your data into a presence/ absence data. In my example below, I create a matrix with some sample counts ranging from 0 to 5. I then use ifelse() to create the presence/absence matrix in which all cells with a count of 0 are coded as 0 (i.e. absent) and all with a count above 0 as 1 (i.e. present). With a matrix of this format, you should be able to continue the tutorial.

I hope this will help!

# generate fake data 
set.seed(1)
my_mat <- matrix(sample(0:5,25,replace = TRUE),nrow = 5)
rownames(my_mat) <- paste("Species",LETTERS[1:5],sep = "_")
colnames(my_mat) <- paste("Sample",LETTERS[1:5],sep = "_")
my_mat
#>           Sample_A Sample_B Sample_C Sample_D Sample_E
#> Species_A        0        2        0        5        0
#> Species_B        3        5        4        1        0
#> Species_C        0        1        4        0        5
#> Species_D        1        2        1        4        4
#> Species_E        4        2        5        4        4

# create absence/presence matrix, 1 denotes present
pres_mat <- ifelse(my_mat == 0,0,1)
pres_mat
#>           Sample_A Sample_B Sample_C Sample_D Sample_E
#> Species_A        0        1        0        1        0
#> Species_B        1        1        1        1        0
#> Species_C        0        1        1        0        1
#> Species_D        1        1        1        1        1
#> Species_E        1        1        1        1        1

[Hira_12]

Thank you for replying.
I used ifelse function as follows:

pres_mat <- ifelse(diss.mat == 0,0,1)

I get the presence absence matrix but my rownames and column names are same.

In the above sample presence absence matrix that i had posted from the tutorial: In the last line it says
rows are species and columns are islands

How can get such type of presence absence matrix according to my data that i have provided in a picture above where i have labelled my rows and columns.

I want row names as

c("Cupriavidus(100)", "Pseudomonadaceae_unclassified(100)",
"Brevibacillus(100)", "Planococcaceae_unclassified(93)", "Rhizobiaceae_unclassified(78)",
"Candidatus_Udaeobacter(100)", "Flavobacterium(100)", "Stenotrophomonas(100)",
"Chitinophaga(100)", "Rhodanobacter(100)", "uncultured(100)",
"Chryseobacterium(100)", "uncultured_ge(99)", "Subgroup_2_ge(100)",
"uncultured(100)", "Pseudomonas(100)", "uncultured(99)", "Enterobacteriaceae_unclassified(97)", "Xanthobacteraceae_unclassified(100)")

and column names as :
c("S1N3d", "S1N6d",
"S1N9d", "S1O3d", "S1O6d", "S1O9d", "S1SB3d", "S1SB6d", "S1SB9d",
"S1Soil", "S2N3d", "S2N5d", "S2N8d", "S2O3d", "S2O5d", "S2O8d",
"S2SB3d", "S2SB5d", "S2SB8d", "S2Soil", "S3N15d", "S3N4d",
"S3N8d", "S3O15d", "S3O4d", "S3O8d", "S3SB15d", "S3SB4d",
"S3SB8d", "S3Soil")

Regards
Hira

[jms]

I suspect that your diss.mat did not actually have the structure that you have shown in your first screenshot and in the linked table. If you load the file that you have shared with us, you should get a dataframe that has the sample names as column names and the genra as the first column. You can then use tibble::column_to_rownames()to set the genra as rownames.

Here is an example for this:

# Create fake dataset with  rownames in first column
set.seed(1)
my_mat <- matrix(sample(0:5,25,replace = TRUE),nrow = 5)
colnames(my_mat) <- paste("Sample",LETTERS[1:5],sep = "_")
my_mat <- as.data.frame(cbind("Species" = paste("Species",LETTERS[1:5],sep = "_"), my_mat))
my_mat
#>     Species Sample_A Sample_B Sample_C Sample_D Sample_E
#> 1 Species_A        0        2        0        5        0
#> 2 Species_B        3        5        4        1        0
#> 3 Species_C        0        1        4        0        5
#> 4 Species_D        1        2        1        4        4
#> 5 Species_E        4        2        5        4        4

# Convert first column to rownames
tibble::column_to_rownames(my_mat,"Species")
#>           Sample_A Sample_B Sample_C Sample_D Sample_E
#> Species_A        0        2        0        5        0
#> Species_B        3        5        4        1        0
#> Species_C        0        1        4        0        5
#> Species_D        1        2        1        4        4
#> Species_E        4        2        5        4        4

If you need help that is more specific to you actual data, it would be helpfull if you could share a representative subset of your data in a copy-paste friendly format. You could for example use the dput() function on your matrix (or like the first 5 rows and columns) and share the output with us.

[Hira_12]

You are right Sir. My diss.mat did not actually have the structure as shown in screen shot.
(Sir, I am giving a smaller part of my data in a copy-paste friendly format as you suggested.)

My matrix that has sample names as column names and the genera as the first column was as follows:

dput(numeric_matrix)
structure(list(S1N3d = c(32802, 28477, 118717, 19454, 38, 11431,
0, 0, 69, 921, 303, 1, 1686, 6802, 1045, 699, 28, 194, 965),
S1N6d = c(11821, 72217, 106281, 4626, 7475, 2112, 3, 1, 9,
515, 25499, 7, 222, 585, 260, 3837, 1446, 107, 221), S1N9d = c(14138,
40757, 57791, 12138, 13595, 5679, 5, 1, 6, 492, 53297, 1,
768, 826, 1542, 2467, 899, 18, 1052), S1O3d = c(51103, 21238,
116748, 10147, 9, 6358, 0, 1594, 244, 706, 3, 1, 953, 3258,
702, 392, 28, 197, 635), S1O6d = c(62970, 15574, 125389,
1433, 127, 2096, 2, 2151, 225, 22701, 23, 7, 256, 885, 233,
311, 1055, 89, 284), S1O9d = c(55195, 7825, 98369, 11735,
2140, 6087, 0, 618, 81, 30320, 10208, 8, 831, 989, 1851,
621, 7027, 67, 1323), S1SB3d = c(61711, 3853, 88559, 8832,
7, 9636, 3, 20, 145, 475, 7, 2, 1133, 4786, 901, 634, 29,
102, 1266), S1SB6d = c(71984, 25757, 130608, 662, 6, 2101,
13, 64, 14, 13661, 9, 16, 236, 770, 238, 534, 1437, 68, 264
), S1SB9d = c(60765, 13045, 63699, 7067, 13, 5771, 4, 38,
12, 18180, 3488, 3, 987, 1056, 1793, 588, 8202, 39, 1539),
S1Soil = c(4, 2, 17, 13317, 3, 16553, 3, 0, 3, 1065, 11,
4, 4719, 8300, 6723, 917, 15, 2, 5883)), class = "data.frame", row.names = c("Cupriavidus.100.",
"Pseudomonadaceae_unclassified.100.", "Brevibacillus.100.", "Planococcaceae_unclassified.93.",
"Rhizobiaceae_unclassified.78.", "Candidatus_Udaeobacter.100.",
"Flavobacterium.100.", "Stenotrophomonas.100.", "Chitinophaga.100.",
"Rhodanobacter.100.", "uncultured.100.", "Chryseobacterium.100.",
"uncultured_ge.99.", "Subgroup_2_ge.100.", "uncultured.100..1",
"Pseudomonas.100.", "uncultured.99.", "Enterobacteriaceae_unclassified.97.",
"Xanthobacteraceae_unclassified.100."))

Then i run the following command:

distances <- vegdist(numeric_matrix, method = "bray")

I get distances as follows:

dput(distances)
structure(c(0.564736701482407, 0.36403669531434, 0.702686441207726,
0.894986835790503, 0.790849593222344, 0.999843796594766, 0.978982622136868,
0.996182385583781, 0.656031231855867, 0.844747458478949, 0.999763337695808,
0.967413029261958, 0.911432057681642, 0.960852572404924, 0.953461762935042,
0.908936675861103, 0.995828766864442, 0.965347250100361, 0.596897763108158,
0.577521718905191, 0.814306902814902, 0.693224848180356, 0.999720252821513,
0.961523290114564, 0.992968943991148, 0.665077521941211, 0.515437214118467,
0.999580410411067, 0.941181361625703, 0.851938895417156, 0.929788184384899,
0.915868944086425, 0.838070635689062, 0.99230930025955, 0.937640651259203,
0.847103573864315, 0.949627309214482, 0.894686047872592, 0.999927169279561,
0.990145662784888, 0.998218274592993, 0.823177728609123, 0.814122955758909,
0.999889652493633, 0.984555033993523, 0.957249032838025, 0.981373159725915,
0.977975921794897, 0.956461098684722, 0.998053052661287, 0.983545198508063,
0.661295469049138, 0.215416414920342, 0.999262108134699, 0.919721399816823,
0.982088030237533, 0.719009005475015, 0.661843859562491, 0.998882194475805,
0.766980889705737, 0.521552163714859, 0.707962826770074, 0.780900498949318,
0.666791388703834, 0.980441668327907, 0.738786305339206, 0.777853283207471,
0.997782137678069, 0.943942652329749, 0.975393253788035, 0.940417433681046,
0.599126104196592, 0.996675616928781, 0.885013066696966, 0.899245209986453,
0.800728663342033, 0.585737947868538, 0.78530943803208, 0.9706947645703,
0.848066223368164, 0.999027366373403, 0.877418373414833, 0.976454132183238,
0.742075736325386, 0.779401513642701, 0.998526681792734, 0.703799535263455,
0.411808786336529, 0.632110886514583, 0.764665584086065, 0.618252074099329,
0.97429665099626, 0.669390568081126, 0.989380530973451, 0.921521997621879,
0.999259001448315, 0.999375545052271, 0.349397590361446, 0.994418132611637,
0.997667020148462, 0.995692187194047, 0.994017946161516, 0.996732511510471,
0.930131004366812, 0.995098403267731, 0.774126534466478, 0.923034975353656,
0.985616684645811, 0.982808022922636, 0.760289961911783, 0.803261666259467,
0.83039190897598, 0.813391877058178, 0.851944996552144, 0.822346368715084,
0.814610190300798, 0.982013267441343, 0.995259246604596, 0.89044289044289,
0.87173585205175, 0.944400481678995, 0.899602385685885, 0.863143631436314,
0.958520072470678, 0.329390892962744, 0.886516853932584, 0.834410943238548,
0.99887748916777, 0.877195592450435, 0.86573793832539, 0.846056516237874,
0.889159902435123, 0.655171150711525, 0.98036010186946, 0.85909747433345,
0.999074253482314, 0.939525415953899, 0.937046364724826, 0.892690685803063,
0.848451583082967, 0.770984125860513, 0.989992638508071, 0.915487391795258,
0.99155476733384, 0.996467304906913, 0.993480245142783, 0.990950226244344,
0.995053423031262, 0.897106109324759, 0.992582702863077, 0.411156612065521,
0.2138557553824, 0.532008248870168, 0.787214068905091, 0.860659618115828,
0.147524085160369, 0.404156619588931, 0.688870774638918, 0.784771699399046,
0.93939601921757, 0.405646573436638, 0.534083992696287, 0.6968466181531,
0.890792158802795, 0.0954038997214485, 0.711287941988064, 0.851384330556257,
0.535690897184021, 0.954867214594518, 0.724388356449789, 0.876632902549773
), Size = 19L, Labels = c("Cupriavidus.100.", "Pseudomonadaceae_unclassified.100.",
"Brevibacillus.100.", "Planococcaceae_unclassified.93.", "Rhizobiaceae_unclassified.78.",
"Candidatus_Udaeobacter.100.", "Flavobacterium.100.", "Stenotrophomonas.100.",
"Chitinophaga.100.", "Rhodanobacter.100.", "uncultured.100.",
"Chryseobacterium.100.", "uncultured_ge.99.", "Subgroup_2_ge.100.",
"uncultured.100..1", "Pseudomonas.100.", "uncultured.99.", "Enterobacteriaceae_unclassified.97.",
"Xanthobacteraceae_unclassified.100."), Diag = FALSE, Upper = FALSE, method = "bray", call = vegdist(x = numeric_matrix,
method = "bray"), class = "dist")

Next, for obtaining an an presence absence matrix I run the following set of commands:

diss.mat <- as.matrix(distances)
diss.cutoff <- 0.6
diss.adj <- ifelse(diss.mat <= diss.cutoff, 1, 0)

Here at this step I do not get the presence absence matrix with the same structure of my numeric_matrix. i.e. sample names as column names and the genra as the first column.

Please tell me what am i missing here to get my desired result. I hope i was able to explain my question more clearly.

Regards
Hira

[jms]

Is there are reasen for not creating the presence/absence matrix directly from numeric_matrix?
The vegdist() function returns a dist object which you can only convert into a species by species matrix, not in a species by site matrix.

[Hira_12]

Thank you very much for replying. This solves the problem.