PAST Lesson 9: Analysis of Similarities (ANOSIM) and Similarity Percentage (SIMPER)
ฝัง
- เผยแพร่เมื่อ 9 ก.พ. 2025
- Paleontological Statistics. Version 4.05 (www.nhm.uio.no.... Here we looked used ANOSIM to determine whether the diatom community profiles were different among sites. If yes, then we used SIMPER to list the most important taxa that contributed to this dissimilarity. For the indices used; Bray-Curtis is often used for abundance data. For absence-presence data (binary data), you may use either Dice or Jaccard.
![One-Way ANOVA [Analysis of Variance] simply explained](/img/n.gif)
EDIT: Ok, I realized that PAST is the actual name of the program you're using. Couldn't tell that at first! And just FYI, the sound is super soft on my computer, despite that not beingn the case for other videos. But I'm still eager to get your opinion on my second question! Do you think that this method (ANOSIM & SIMPER) is justifiable for partly non-abundance, non-species data (some numeric, some ordinal and dummy-coded)? Am looking to compare contribution of forest habitat attributes (some of which are tree-species based, but some not) to differences between where competing predators are foraging. I'm doing it in R but would love to verify it with a different program to ensure my results are similar. But I also kinda just want to make sure it's ok to even do because no one on my committee knows. >
Simper usually done after Anosim if significant. For Anosim, i isually use ratio or count data. Although Anosim ranks the data, i don’t think ordinal data is suitable as its distance is not comparable
@@LeeChoonWeng Thank you so much for the reply! Very interesting.
Hi Sir, can I used data before recompute the "Or pool all groups" ? Seems like my data wont recompute after i tick the "Or pool all groups" box. Your reply is highly appreciated. Tq!
If in simper, and you pool your groups, the default index used is bray Curtis and cannot be changed. Therefore it will not allow recompute.
Sorry but I couldn’t access your earlier question
Thanks
Hi, I am trying to run a ONE-WAY ANOSIM on my data to see if there is a similarity among treatments and once I try to run, an error pops up saying, at least two groups are required. How do I go about this?
Did you set your first column to groups?
Hello. Could you please elaborate "a minimum of 2 rows is required for each sample/island"? Does that mean each sample/island is surveyed twice?
Yes. Apparently when I tried with one row, an error message popped up 'requiring two rows per group'.
@@LeeChoonWeng I only surveyed the islands (group) A and B once and divided their abundance into 2 rows so each island (group) has 2 rows. PAST was able to analyzed it.
I also tried putting the second rows of each island (group) as 0 abundance, PAST was able to analyzed it too but a totally different result that doesn't reflect the abundance table was produced. I believe the 1st method like the one you did in the video is correct.
@@syunsn7230 I remembered you said you measured 5 islands. I will probably calculate the alpha and beta diversities. Then use cluster analysis to check for similarities with Bray Curtis coefficient.
@@LeeChoonWeng Those 5 islands are from a different set of locations. Yes, I've calculated the diversity indices at every site and had run the cluster analysis, but it was Jaccard instead of Bray Curtis. I will run Bray Curtis too. Thank you for your input.
@@syunsn7230 Good luck
Thanks a LOT. 🙏🏽
Glad the video helped
Thank you so much
How can i save my data please?
Sorry for the late reply. You may save your file as a dat file. For results of analysis, I usually copy and paste into excel. For graphs, you may export as svg or png files
can you give us the data source?
Any reason why?
@@LeeChoonWeng wanted to read the paper so as to cite the sources
@@akshayasrini6542 Lim JH, Lee CW (2017). Effects of eutrophication on diatom abundance, biovolume and diversity in tropical coastal waters. Environmental Monitoring and Assessment 189: 432
Hope this helps
@@LeeChoonWeng Thanks a lot!