How to identify gaps or critical dyads on the lower level?

[meefen]

The motifr documentation does provide useful guidance on how to identify gaps and critical dyads. See example code below:

plot_critical_dyads(ml_net, "1,2[I.C]",
  level = -1,
  subset_graph = "focal", cutoff = 4, label = TRUE
)

However, I could not get it to find gaps or critical dyads on the lower level (e.g., level 0). When I tweak the level parameter, errors are often reported.

Is there a way to identify gaps or critical dyads on the lower level? Thanks.

[marioangst]

Thanks for writing in!
Given that you did not provide a reprex and no error messages, I can somewhat only guess what your trouble is.

If you could provide a reprex, that would be great so we can investigate further. Also check that the motif even exists in that configuration.

Could it be that you might be helped by specifying the levels in the motif signature? see here: https://marioangst.github.io/motifr/articles/motif_zoo.html under "A word of caution on counting: Position matching". So possibly "1:1,2:0[I.C]" would work. When in doubt, use show_motif to see that your motif label does what you expect it to do.

[meefen]

Thank you @marioangst for the suggestion. I was able to achieve what I hoped for by specifying the levels.

See sample code below:

plot_gaps(ml_net, "2:1,2:0[III.D]") # return level 1 gaps
plot_gaps(ml_net, "2:0,2:1[III.D]") # return level 0 gaps

I still wonder how the level parameter can be used. It specifies the "Focal level for gap analysis" and isn't it suppose to achieve the same outcome while avoiding changing the motif? Please advise.

[marioangst]

@meefen the focal level argument is only useful when you have motif where the location of the gap is not apparent. Eg. in a four-cycle, where a gap can be on either level. For triangles, the gap can only be on the level where there are two nodes. Does this make sense?

[meefen]

Thanks for the reply. Yes it does make sense.

For four-node motifs of two nodes for each level, is there a way to capture gaps on a specific level?

[marioangst]

Yes, in this case, the level argument should work to specify the location of the gap.

[meefen]

It didn't work for me even with the example dataset. I do not have access to my computer right now and can post sample code if needed.

[marioangst]

Oh that would be great, please give us a reprex so we can investigate!