cogent3 / ensembldb3

ensembldb3

WARNING This tool is being replaced by EnsemblLite. EnsemblLite lite is not yet ready for use, but should be in early 2024. Thanks for your patience!

Documentation

ensembldb3 documentation is available at read the docs.

ensembldb3's Issues

download should use CHECKSUMS

The ensembldb3.download.Download class should check that all downloaded files match the value in Ensembl's provided CHECKSUMS file.

Add a function based on the following code that is called by Download.

def get_invalid_files(path):
    """return files with invalid checksums"""
    checkpath = (path / "CHECKSUMS")
    failed = []
    with open(checkpath) as infile:
        for record in infile.readlines():
            record = record.split()
            filepath = path / record[-1]
            if not filepath.exists():
                continue

            result = exec_command(f"sum {filepath}")
            result = result.split()
            if result != record[:2]:
                failed.append(filepath)
    return failed

Where do gene symbols come from for human ensembl genes?

I am trying to figure out where ensembl stores the symbol for a gene in its MySQL database (schema). Forgive me for opening an issue here. My excuse is that:

• I don't find email lists very effective for technical discussion or discoverability, so I haven't yet tried the [email protected] mailing list
• The official REST API does provide a display_name field, e.g. display_name: BRCA2 for https://rest.ensembl.org/lookup/id/ENSG00000139618. But I want to know how this is retrieved from the MySQL database and prefer reading Python to Perl.

From the ensembldb3 docs:

>>> gene = human.get_gene_by_stableid(stableid='ENSG00000139618')
>>> print(gene)
Gene(species='Homo sapiens'; biotype='protein_coding'; description='BRCA2, DNA repair...'; stableid='ENSG00000139618'; status='KNOWN'; symbol='BRCA2')

Note symbol='BRCA2'. Now I'd like to figure out where in the database this is coming from.

improve installation performance

At present, installation status of a db is identified via existence of an empty ENSEMBLDB_INSTALLED file.

If installation is interrupted, for some reason, then the entire db is dropped and every table reinstalled. Instead, this file should be used to record successful installation of each table.

That means, the status check needs to read the contents of this file.

external listing of code quality gives us a D!

Code quality measures does not give us a very good score, but much of these are readily fixed.

should handle files already decompressed

If both the compressed and decompressed files exist, unlink the decompressed and gunzip again to ensure it's fully decompressed.