How to ask panacus just openness information

[AndreaGuarracino]

Hi, what command should I run on a GFA to get information about opening the pangenome? I'm a bit lost about old/new documentation.

The plan is to get such info on thousands of GFA files representing HPRCv2 pangenome partitions.

[AndreaGuarracino]

I am interested in the information at the top!

[heringerp]

Hi Andrea,
unfortunately currently there is no command to just get the openness of the pangenome, the function is currently only part of the HTML file. However, I should be able to quickly add that information to the growth command. Let me get back to you as soon as that is implemented.

[heringerp]

So I've implemented this feature now, but it is still only part of the feature branch regions. If you build this version of panacus, you can get just the alpha/growth values with

panacus growth --alpha partition.gfa | grep "alpha"

[Matteo-Paini]

Hi @heringerp, thanks for this implementation! I tried it and it worked.

Hower, I have a couple of doubts:

• I'm not really sure how to interpret that "alpha" value, may I ask you an explanation about it?
• Secondly, I'm really interested in those graphs posted above: I didn't understand if they can be generated with panacus and, if so, how?

Thanks in advance for your reply!

[heringerp]

Hi,

1. In general α ≥ 1 means that a given pangenome is closed, i.e., there is a finite amount of features (nodes, edges, bps in the case of panacus) in your population. At some number of individuals you will have seen everything there is. α < 1 means the pangenome is open, no matter how many individuals you sample, you will always find new features. For the original definition see Tettelin, et al. Comparative genomics: the bacterial pan-genome. 2008. https://doi.org/10.1016/j.mib.2008.09.006. Interpreting much more from this value is difficult. However, we are currently working on this value, what it can say in more detail and what alternatives there are, so expect more in a couple of months.
2. Yes, the graphs above can be generated with panacus, but only with the code on the regions branch (https://github.com/codialab/panacus/tree/regions) using something like this:

- graph: ../simple_files/graphs/pggb/chr22.hprc-v1.0-pggb.gfa
  grouping: !Custom ./chr22_group_samples.tsv
  nice: false
  analyses:
    - !RegionalGrowth
      reference: "grch38#chr22"
      count_type: Node
      log_windows: false
      reference_subset: easy_chr22.bed
      merge_small_windows: true

However, please not that I am currently developing on this branch so the features there will change suddenly and without any warning.

Best regards,
Peter

[Matteo-Paini]

Hi Peter, thanks for the detailed and fast reply! I was able to generate the plots and now I better understand the meaning of α.

I still have just a couple of final (I hope) questions:

1. So the value of α is like the Growth Curve analysis which uses the Heaps Law? And can't wait for the next releases with the updates on this value!
2. In the html output file for the plot there are two variables: "Growth" and "Hapax-Ratio". This last one is not very clear to me, can I ask you an help to interpret it correctly?

I'm sorry for all those questions, I'm kinda new in this field and your help is really appreciated!

[heringerp]

1. Yes, this is indeed the α value of the Heaps' Law
2. Growth is the α value, while Hapax-Ratio is the amount of singleton features (hapax legomena in the field where Heaps' Law is coming from) divided by the total number of features