cdot gene biotype is a list

[davmlaw]

Problem:

cdot JSON used to have gene biotype as a string, but it has changed to being a list

In create_biotype_regions_array this is returned:

    def get_biotype(gene):
        if gene.biotype in other_biotypes:
            return "other"
        elif gene.biotype == "misc_RNA":
            if gene.name and gene.name.startswith("RNY"):
                return "yRNA"
        return gene.biotype

Which returns a list, this eventually causes a crash of:

        length_counters[length][read_region] += 1

unhashable type list

Fix:

The difficulty here is that we are assuming 1 biotype, while we may have to pick

The reason we are passing gene into get_biotype() rather than just relying on biotype in the loop:

    for biotype in other_biotypes + interesting_biotypes:

Is because we have special case code handling yRNA

We can possible break this out to only handle the biotype

[davmlaw]

So I think the trick is to pass in the

    def get_biotype(biotype, gene):
        if biotype in other_biotypes:
            return "other"
        elif biotype == "misc_RNA":
            if gene.name and gene.name.startswith("RNY"):
                return "yRNA"
        return biotype


    for biotype in other_biotypes + interesting_biotypes:
        genes = reference.genes_by_biotype.get(biotype, [])
        for gene in genes:
            for t in gene.transcripts:
                for exon in t.get_features("exon"):
                    regions[exon.iv] = get_biotype(biotype, gene)
                    

I was worried that genes by biotype wouldn't have handled this string to list change, but looks good?

# reference._gene_id_lookups
            if gene_version:
                gene_transcripts[gene_version].add(transcript_id)
                # In cdot 0.2.20 onwards gene version will have biotype of any transcripts, but earlier this wasn't so
                for biotype in tdata["biotype"]:
                    gene_version_by_biotype[biotype].add(gene_version)

[davmlaw]

Should make a unit testing with latest cdot version