Roadmap

library(fsmc)

fsmc Roadmap

Improvements

• add fluxes to species edges
• Add warning if there are unexpected values in the diagonal (!=0)
• Throw insightful error when input wrong for MF

New Methods

• Build and document the visualisation method for the alignment
• Function to identify critical/overlapping pathways
• Create a wrapper function around MicrobiomeFunction that allows for easy import from relevant sources, including: SBML (Cobra CarveMe) JSON (COMETS, Escher) YAML (Misosoup). This needs to: accept factor variables for species and metabolite, allow for flexible variable naming, generate community name if missing
• Write set functions for all slots
• Write output methods to different formats

Visualisation

MicrobiomeFunction

• Graph Plotting method
• Plotting functions for EDA with info on the different metabolites/reactions ### MicrobiomeFunctionSet
• Plotting functions for EDA with info on the different communities ### MicrobiomeFunctionAlignment
• how to visualise the alignment? Read over the cooccurrence paper again and see which plotting solutions from them might make sense.

Methods - EDA

MicrobiomeFunction

• accessor method for produced/consumed by a species, should be implementation of getMet with additional arg = “consumed”/“produced” and additional one for species.
• calculate effective fluxes
• find longest pathway
• methods for common graph metrics ### MicrobiomeFunctionSet
• Same as for MF object, but additional argument “community” to select one within the set.
• show method includes number of communities, average/min/max size in terms of taxonomy and metabolites ### MicrobiomeFunctionAlignment
• show levels of alignment
• Compare connected vs unconnected alignment
• get scores of different alignment
• get alignments including certain species or metabolites

Alignment Methods

• Make the methods quicker by not including alignments that are under the cut off limit
• write a search function that either takes an MF object or the name of an MF object in an MFS to search the closest relative to.
• Toggle connected and disconnected alignment
• Improve speed
• Warning if aligning MF with same ID
• minimum flux of production is what is used in the alignment
• Percent Identity (improve speed )
• return (dis-) connected modules
• connectedness of the alignment (as compared to the connectedness of all reactions?)
• centrality for each edge in the alignment
• make the method so that it takes a min number of reactions to be in the alignment, same argument for threshold cutoff as with the depth
• if an alignment is not completley connected, evaluate the connectedness of each level of alignment.
• pairwise vs multiple alignment?
• Return metric on the average length of the reactions in the alignment?
• peripheral vs central communities in the alignment? This would have to be solved with making a big graph of all communities and then giving each node groups for the communities it’s in. Then for each group the average centrality would be a good measure.
• How would we introduce gaps into the alignment? This could be most easily done by introducing placeholder metabolites/species. What is the best way to go about this? Matrix multiplication or graph representation?
• introduce option to align with or without gaps with an argument gaps = TRUE/FALSE
• What is the maximum connected breadth that we can achieve with a set of communities? –> This is what gaps will be most useful for.

General Ideas

• Can performance be improved by switching to use a databases?
• Reread hypergraph paper and consider implications

Documentation

• Document the different types of input data from various sources such as FLYCOP, MisoSoup, etc.
• What are the differences in the formats?
• How can each be imported into fsmc?
• Do standard formats exist?
• Which formats should fsmc support in the future?
• What does data from carvme look like? How does it differ from Misosoup/Flycop

Long Term

• Chemical substitution score for metabolites
• clusterMet method to do the clustering of the metabolites and get an overview over how similar the metabolites in the community are.