cruzDB coordinates vs UCSC coordinates

Question

[nkinney06]

I am confused by a subtle difference in the cruzdb coordinates and UCSC coordinates. For those unfamiliar cruzdb is a python package for accessing the UCSC databases. Suppose I am interested in exon 3/7 for the gene TAF3. This exon extends from base pair 7963920-7965742 on chromosome 10. You can be sure of this by zooming in on the two ends of the exon in the UCSC genome browser:

https://genome.ucsc.edu/cgi-bin/hgTracks?db=hg38&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr10%3A7963919-7963922&hgsid=651863061_L21GEzydvXVo8gitnxDdfRcTc4hi

https://genome.ucsc.edu/cgi-bin/hgTracks?db=hg38&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr10%3A7965740-7965743&hgsid=651863061_L21GEzydvXVo8gitnxDdfRcTc4hi

However; cruzdb appears to agree with the start of the exon but is off by one base at the stop:

import cruzdb
from cruzdb import Genome

g = cruzdb.Genome('hg38')

# case 1 only picks up the cds which appears to be the exon range according to cruzdb
chromosome = 'chr10'
start = 7963920
stop = 7965741
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 2 also pick up the intron on the left
chromosome = 'chr10'
start = 7963919
stop = 7965741
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 3 also picks up the intron on the right but this is the exon range according to UCSC
chromosome = 'chr10'
start = 7963920
stop = 7965742
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

I post this at the risk of looking really dumb, but I cant risk annotating hundreds of regions and be off by one base. Thanks!

[nkinney06]

Update: for anyone that uses cruzdb you may want to look into this unsolved issue. I received this comment from the developer:

probably, these comparisons:
https://github.com/brentp/cruzdb/blob/master/cruzdb/models.py#L81 need
to use >= instead of >. If you verify that, I'll make the change or
accept a PR.

Answer 1

[max]

UCSC coordinates are 0-based and half-open internally (like cruzDB) but 1-based and closed in the html user interface.

http://genomewiki.ucsc.edu/index.php/Coordinate_Transforms

Answer 2

[nkinney06]

I'm surprised this question has received little attention as I am quite convinced there is an inconsistency in the cruzdb coordinate system; running the following code should be convincing enough.

import cruzdb
from cruzdb import Genome

g = cruzdb.Genome('hg38')

#####################################################################
# first an example of the error on the negative strand
#####################################################################

# case 1: this query says the only features are intron 
chromosome = 'chr9'
start = 35812101
stop = 35812102
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 2: this query says the features are intron and cds
# therefor we can conclude that 35812103 is the first base in the cds
chromosome = 'chr9'
start = 35812102
stop = 35812103
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 3: this query says the features are intron and cds which does
# not make since because we concluded that 35812103 was in the cds
chromosome = 'chr9'
start = 35812103
stop = 35812104
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

#####################################################################
# an example of the error on the positive strand
#####################################################################

# case 1: both bases in the cds
chromosome = 'chr7'
start = 128074502
stop = 128074503
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 2: cds and intron - conclude 128074501 is intron
chromosome = 'chr7'
start = 128074501
stop = 128074502
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features

# case 3:  cds and intron - inconsistent with the above
chromosome = 'chr7'
start = 128074500
stop = 128074501
gene = g.bin_query('refGene',chromosome,start,stop).all()
features = [x.features(start,stop) for x in gene]
print features