Missing chromosomes about mutalyzer2 HOT 4 OPEN

https://github.com/alexpenson/scripts/blob/master/chr_ncbi_to_ucsc.sed

But there must be a better source (where we originally got them from?).

from mutalyzer2.

This is really a mess. Looks like we can sort of get what we want by first going to UCSC genome browser and selecting from ctgPos2 (GRC Map Contigs) everything that starts at 0 (i.e., is the same as the complete chromosome it is located on):

$ mysql --user=genome --host=genome-mysql.cse.ucsc.edu -A -D hg19
mysql> select * from ctgPos2 where chromStart = 0;
+----------------------+---------+-----------------------+------------+----------+------+
| contig               | size    | chrom                 | chromStart | chromEnd | type |
+----------------------+---------+-----------------------+------------+----------+------+
| HSCHR17_CTG1         |  296626 | chr17                 |          0 |   296626 | F    |
| HSCHR1_RANDOM_CTG5   |  106433 | chr1_gl000191_random  |          0 |   106433 | F    |
| HSCHR1_RANDOM_CTG12  |  547496 | chr1_gl000192_random  |          0 |   547496 | F    |
...

Now go to the GRC website, click on the assembly accession you want, and click Download the full sequence report:

In that file, lookup the contig we got from the UCSC database, (e.g., HSCHR1_RANDOM_CTG5) and observe the corresponding refseq accession number (e.g., NT_113878.1).

from mutalyzer2.

@p.e.m.taschner What do you think? Is there an easier way to get the RefSeq accession numbers and UCSC chromosome name couplings?

Here's an example of what we already have.

from mutalyzer2.

After discussing this:

The mappings on these contigs that have no chromosome definition are bogus anyway, since there is no way to use them in the position converter (Mutalyzer will always first try to get the corresponding chromosome definition). So we decided to discard them for now.

At some point we should decide what to do with all contigs that are not part of the primary assembly. But currently, our user interface is not really suited for using them, so that needs some more thinking. It may really confuse users if we start adding more mappings on these non-primary assembly contigs.

By the way, we only have transcript mappings on the non-primary assembly contigs from our old UCSC imports. The current NCBI mapview import discards them completely. We're not fixing this before we decided how to handle and present them to the user.

from mutalyzer2.