I’m not sure if this is a bug or just something I’ve done wrong so please move to another forum if necessary.
Below is the subsampled avg Yue & Clayton theta on 4 communities (i.e. this file is stability.an.thetayc.0.03.lt.ave.dist when working through the MiSEQ SOP)
S59 0.979511 0.944713
S60 0.976537 0.450679 0.747765
When I use this file as the input for heatmap.sim the image looks fine for most of the comparisons (i.e. dark brown/deep red on the similarity scale because the communities are mostly not very similar). However, for the S58-S60 comparison (0.45 dissimilarity thetayc which should be 0.55 similarity in my heatmap) the colour is extremely bright red, close to 1.0 on my scale bar in that image (at least well above 0.8).
Is there any way I can fix this because if I understand correctly, the heatmap.sim image should reflect the data in the table above but it definitely doesn’t appear to?
I have done various other heatmap.sims (for other communities) and never encountered this (well, not as visually obvious as this case).
I can send through the svg file if necessary.
Mothur adjusts the colors to map the closest similarity to red and gradates it to black for the least similar.
Sure, but shouldn’t the scale bar in the image still match the values in the raw data?
Or are you saying that the most similar communities in my particular comparison will appear as 1.0 bright red (even if they are only 0.55 similar) and the least similar communities in my particular comparison will appear as black, 0, (even if they are 0.1)?
Could you post the image for me?
This should be publicly accessible but please let me know if not
The one I have underlined (thetayc dissimilarity 0.45) is the one I am concerned about in the heatmap (bright red).
Any ideas anyone?
I’ve gone through some other recent heatmaps from my other datasets and it looks to me like the least similar communities are scaled black while the most similar are scaled bright red, instead of the colours actually reflecting the similarity values…
If there is a way around this please let me know as I really like the heatmaps and think they can show the data much better than a table listing the similarity (or dissimilarity) index will…