Why one protein appears more once in ucsc knownGene? And why some UniProt proteins are missing in knownGene?

[Peng Yu]

Hi,

The following mysql code shows that many proteins appears more than
once in knownGene.

~~~
mysql> select count(*) from knownGene where proteinID!="";
+----------+
| count(*) |
+----------+
| 41735 |
+----------+
1 row in set (0.32 sec)

mysql> select count(distinct proteinID) from knownGene where proteinID!="";
+---------------------------+
| count(distinct proteinID) |
+---------------------------+
| 34117 |
+---------------------------+
1 row in set (0.23 sec)

mysql> select proteinID from knownGene group by proteinID having
count(*) >=2 limit 5;
+-----------+
| proteinID |
+-----------+
| |
| A0ELI5 |
| A0JNT0 |
| A0JNY8 |
| A0JNZ2 |
+-----------+
5 rows in set (0.26 sec)
mysql> select * from knownGene where proteinID="A0ELI5";
+------------+-------+--------+----------+----------+----------+----------+-----------+-----------------------------------------------------------------+-----------------------------------------------------------------+-----------+------------+
| name | chrom | strand | txStart | txEnd | cdsStart |
cdsEnd | exonCount | exonStarts
| exonEnds
| proteinID | alignID |
+------------+-------+--------+----------+----------+----------+----------+-----------+-----------------------------------------------------------------+-----------------------------------------------------------------+-----------+------------+
| uc009pvp.1 | chr9 | + | 57556375 | 57597969 | 57561204 |
57596302 | 7 |
57556375,57561187,57563754,57574992,57587696,57592391,57595967, |
57556506,57561368,57564074,57575328,57587850,57592609,57597969, |
A0ELI5 | uc009pvp.1 |
| uc009pvq.1 | chr9 | + | 57556375 | 57597969 | 57575057 |
57596302 | 6 |
57556375,57561187,57574992,57587696,57592391,57595967, |
57556506,57561368,57575328,57587850,57592609,57597969, |
A0ELI5 | uc009pvq.1 |
+------------+-------+--------+----------+----------+----------+----------+-----------+-----------------------------------------------------------------+-----------------------------------------------------------------+-----------+------------+
2 rows in set (0.07 sec)
~~~

The following shows that UniProt has more mouse proteins than the ones
in knownGene.
~~~
~$ wget -qO- ftp://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/proteomes/MOUSE.fasta.gz
| gunzip | grep '^>' | wc -l
50798
~~~

Does anybody know why for both these questions?

--
Regards,
Peng

[Luvina Guruvadoo]

Hi Peng,

One of our engineers has this to say: The UCSC Genes process doesn't use
the UniProt proteins for its build so there is no guarantee that there
will be a model for every protein, and since it creates more than one
model for a particular locus if there are different transcripts with
that locus, there is no guarantee that there won't be more than one
transcript for a given UniProt protein. The many to many mapping is
expected given the algorithm, which is documented on the UCSC Genes
track description page
(http://genome.ucsc.edu/cgi-bin/hgTrackUi?db=mm10&g=knownGene).

I hope this helps. If you have further questions please feel free to
contact the mailing list again at gen...@soe.ucsc.edu.

---
Luvina Guruvadoo
UCSC Genome Bioinformatics Group