Could alternative methods for BLAST be used?

[Budjavac]

I was wondering if there was a way of incorporating an alternative BLAST method into this pipeline. E.g. would it be possible to, instead of using the map_genes.sh command to perform BLAST between each pair of transcriptomes or proteomes via running a DIAMOND alignment via orthofinder?

[atarashansky]

As long as you can generate a homology graph, then any mapper can be used. There's an optional parameter to the SAMAP(...) constructor called gnnm which accepts a custom homology graph. The docstring is a bit out of date, but it's a tuple of (homology_graph, genes, gene_species_dictionary):

• homology_graph is a scipy sparse adjacency matrix with edges connecting mapped genes.
• genes is a numpy array of genes corresponding to each row/col of the homology graph.
• gene_species_dictionary is a dict with species IDs as keys and their respective genes (stored as numpy arrays) as values. Each gene ID in genes and genes_species_dictionary should be prefixed by the species ID (e.g. hu_{GENE_ID/SYMBOL}). Passing this in will override SAMAP's requirement to have the BLAST output available.

[Budjavac]

Thank you! Closing the thread