Introduction
The purpose of the Uprobe project is to provide the community with an efficient means of experimental access to large-insert cloned DNA (ie, BAC clones) from the full spectrum of vertebrate genomes for which genomic BAC libraries are planned or are currently available. This is being accomplished by technology development aimed at improving the ability of 'universal' hybridization-based probes to identify genes/regions of interest from multiple species, and then the dissemination of these improved technologies through the creation of:
- pre-computed whole-genome sets of universal probes,
- on demand design of universal probes for screening nonhuman primate libraries
- custom universal probe design from DNA sequence alignments provided by the user.
A more general introduction to the Uprobe project and details related to the above mentioned resources are provided below.
Also note that you can also download whole-genome probe sets , computer programs , and experimental protocols from this website.
Pre-computed whole-genome probe sets are available for:
All Mammals.
AUG_2003_b1 is based on human-mouse alignments and has been experimentally validated.
OCT_2003_b2 is based on human-mouse-rat alignments and has been experimentally validated.
FEB_2004_mammals_1 was created by combining the approaches used to build the first two probe sets and has been experimentally validated, and is the current recommended and default probe set for screening mammalian genomic libraries.
JUN_2005_mammals_2 was created by enhancing the FEB_2004_mammals_1 with new probes based on human-mouse-rat-dog-chicken alignments and is the current recommended and default probe set for screening mammalian genomic libraries.
Rodents.
APR_2005_rodents_1.1 was designed with a new algorithm, nsoop_v2, using mouse-rat-human-dog alignments specifically for screening rodent libraries and is currently in the process of being experimentally validated. This set replaces OCT_2004_rodents_1.
Carnivores.
APR_2005_carnivores_1.1 was designed with a new algorithm, nsoop_v2, using dog-human-mouse-rat alignments specifically for screening rodent libraries and has been experimentally validated. This set replaces JAN_2005_carnivores_1.
Marsupials.
JUN_2005_marsupials_1 was designed from human-opossum alignments and is the recommended probe set for screening marsupial libraries.
All Birds and Reptiles.
MAR_2004_birds/reptiles_1 is based on chicken-human alignments and has been experimentally validated.
On demand universal probe design for nonhuman primates:
On demand universal probe design for Apes and Old world monkeys is based on near identical sequences from human-chimpanzee-rhesus monkey whole-genome alignments and is currently being experimentally validated.
Similar resources are currently being developed for 3 clades of nonhuman primates: New world monkeys, Simians and All primates.
Custom universal probe design:
Custom universal probe design can be performed on DNA sequence alignments of 2 or more species
provided by the user. A step-by-step tutorial for the custom probe design process is provided here.
The Uprobe project is currently
funded by a grant from NCRR (R24RR022239). Past funding was provided through
the NIH BAC Resource Network by grant (U01MH068185). Comments are welcome and
can be directed to James Thomas (jthomas@genetics.emory.edu). Below is an overview of goals and rationale behind the uprobe project, a description of what universal probes are,
how they are designed, details on specific whole-genome probe sets and updates
and changes to the uprobe site. Comparative
sequencing is a particularly powerful tool for inferring function from genomic
sequence. Thus, sequencing the same region in multiple species simultaneously
would provide a valuable method for interpreting all genomic sequence as it is
generated. Just as important is access to cloned DNA containing orthologous genes and regions in multiple species
catalogued in the sequence. Such access would establish a new resource that
could be used for comprehensive functional analysis of both coding and
non-coding sequence. This new resource would consist of a series of gene
alleles generated not by mutagenesis within a species, but by the divergence of
sequence over evolutionary time (ie 'evolutionary
alleles'). These 'evolutionary alleles' could then be used to experimentally
dissect the function of genomic sequence in cell culture or a transgenic model.
At this time, there is no published method for supplying the templates
necessary for this type of sequencing or functional studies. Since multiple
species sequence comparisons and physical mapping among vertebrates will be a
tremendous resource for the functional annotation of the human genome and a
starting point for experimental analysis, robust and effective means of
generating such clone resources in a targeted manner would be of immeasurable
value to the research community. BAC
libraries offer the means to selectively isolate a specific region of a genome
or to support whole-genome mapping and sequencing efforts. As such, both the
NHGRI (RFA-HG-01-002) and NSF (NSF 01-145) have made funding commitments to
establish a large resource of BAC libraries from a diverse set of species. In
fact, nearly 100 vertebrate BAC libraries are currently available to the public
(Fig. 1) . This new BAC library resource will provide
a source for comparative sequencing and functional studies across a wide range
of vertebrates. Critical to the future utility of these BAC libraries will be
the availability of efficient and reliable methods for screening these
libraries and assembling high-quality BAC maps that can be used by the entire
biomedical research community. This is especially important for individual
researchers who do not necessarily have the experience, technology or
production needs of a genome center. Species-specific
sequence resources traditionally used for isolating and constructing physical
maps of regions of interest are not available for most vertebrates. That is, ESTs or other random genomic sequences. Thus, the
traditional route of building physical maps for a region of interest using
species-specific markers would not be possible for many of the species shown in
Fig 1. The goal of the first phase of the Uprobe
project is to combine the principles of traditional comparative mapping with
existing methods for screening BAC libraries to develop an efficient, practical
and reliable experimental strategy for assembling BAC maps from a diverse set
of vertebrates. Specifically, this proposal aims to remove the limitation of
species-specific resources necessary for BAC library screening by the design
and testing of universal overgo probes that can be
used on single or multiple BAC libraries and will be accomplished by
identifying evolutionarily conserved sequences between species such as human
and mouse, for which there is genomic sequence available. As a result, this
would provide an experimental and computational infrastructure aimed at the one
essential part of BAC library screening not yet standardized, probe design. The
strategy and methodologies proposed here would also dramatically reduce the
cost of building physical maps by increasing the potential mapping throughput
and decreasing the cost of marker reagents while maximizing the ability to
compare genomic maps of divergent species. Through the establishment of a
public database of universal probes, individual researchers would have a key
resource necessary to construct physical maps (independent of whole-genome
efforts) in their region and species of interest that would otherwise be
difficult to build. Small clone-based physical maps assembled by individual
researchers would also complement whole-genome efforts through the direct
integration of both types of maps via the specific clones of interest.
Therefore, it is anticipated that this scalable methodology would greatly
facilitate the use of future BAC libraries by individual laboratories and
genome centers alike, and thus be a widespread means of using the power of
comparative genomics for both community's specific research goals. The ongoing goals of
the Uprobe project are as follows: Modern genomic tools and resources will be critical for
ongoing and future nonhuman primate research and whole-genome sequencing
efforts are underway for a limited number of nonhuman primates. Unfortunately,
because of the sequencing strategies employed and associated costs, nonhuman
primate genome assemblies will have hundreds-of-thousands of gaps and will not
provide a definitive reference sequence like the finished human genome. Moreover,
whole-genome sequences will not yield direct access to the experimental tools
necessary to exploit this extensive genetic information. Bacterial artificial
chromosome (BAC) libraries and clones are a proven and valuable genomic
resource for the experimental utilization and functional characterization of
genomic sequence and are currently available for eighteen species of nonhuman
primates. Input from representatives of the nonhuman primate and biomedical
research communities revealed strong support for a resource to facilitate
access to these nonhuman primate genomic libraries (see appendix). The goal of
this proposal is to develop a resource that will provide an effective and
reliable means for the primate and biomedical research communities to isolate any
specific gene or region of interest from one or all nonhuman primate BAC
libraries. To do so, a web-based tool will be developed for the custom design
of universal hybridization probes that can be used for the isolation of nonhuman
primate BAC clones. Universal hybridization probes are a proven technology for
the efficient targeted isolation of BAC clones from multiple species in
parallel. However, this methodology has not been optimized for use in nonhuman
primates. This proposal will establish those optimal parameters and provide
them as preset values for the custom design of nonhuman primate universal
probes by the public. This custom universal hybridization probe design website
will therefore facilitate access to the full spectrum of genetic diversity
captured within all current and future nonhuman primate genomic libraries
independent of, and as a complement to, whole-genome sequencing efforts. As a
result, this proposal will yield an important avenue by which individual researchers
can readily import nonhuman primate genomic clones into their own laboratories
and experimental paradigms. The aims of this resource proposal are: Aim
1. Develop a public resource
for the custom design of universal hybridization probes for isolating nonhuman
primate genomic clones. A robust universal probe design pipeline used
previously in the creation of whole-genome probe sets will be adapted for the
custom and on demand design of universal probes for isolating nonhuman primate
genomic clones from specific genes and regions of interest. A web-based
interface to this custom probe design pipeline will allow the public to design
probes from a series of default settings catered to the efficient isolation of
genomic clones from four specific clades of primates
(1. all primates, 2. simians, 3. new world monkeys and 4. old world monkeys and
apes), or to design probes from their own sequence data. Aim
2. Experimentally validate
the nonhuman primate custom probe design resource. Small sets of universal
probes designed from each of the four ‘default’ nonhuman primate universal
probe design options will be selected for use in a small-scale targeted
comparative mapping and sequencing project for experimental validation of the
resource, and to provide a real world example of how to use the resource and the
data it can produce. What
are universal probes? The
concept behind the use of universal probes is very simple. If a sequence is
conserved between two divergent species, then it is likely to be conserved in
other species as well. For example, if a sequence is conserved between human
and chicken, then it is likely it will be conserved among all mammals and all
birds. Thus, a single sequence can act as an effective probe for screening
genomic BAC libraries from many birds and mammals and alleviate the need for
generating species-specific probes for every genomic library to be screened. In
addition, since a single probe is used to screen multiple species, screening of
BAC libraries can be done in parallel with identical hybridization and washing
conditions. In the preliminary data for this project, universal probes designed
for screening placental BAC libraries were designed based on sequence
similarity between human and mouse sequence. These probes were tested and found
be effective at isolating BAC clones from a set of placental mammals (cat, dog,
cow, pig, rat, baboon and chimpanzee) (Thomas et al, Parallel Construction of Orthologous Sequence-Ready Clone Contig
Maps in Multiple Species. Genome Res, 2002.
12:1277-1285). The
specifics of the process are illustrated in the figure below. The probes
themselves are called 'overgo' probes and are
comprised of two complementary 22-mers that overlap by 8-bp and are
radioactively labeled with a klenow fill-in reaction
with dATP and dCTP. Overgo probes were developed by John McPherson and have
been used extensively by large and small labs alike to screen genomic
libraries. The specificity and uniform design parameters allows one to
hybridize groups of probes together. Because primers are cheap and the
radioactivity used for labeling a given probe is minimal, we strongly recommend
(when feasible) the use of multiple probes from a given region be used to for
library screening versus a single probe. We aim to design probes that will have
at least a 50% chance of success in a given species,
therefore by using multiple probes, the likelihood of identifying clones of
interest with one hybridization is maximized. When large regions are targeted
for isolation, spacing of the probes every ~30 kb has proven very successful.
This basic process is being used to design whole-genome probe sets for clusters
of species, such as placental mammals, birds, reptiles and sub-groups of fish
using whole-genome alignments. Figure 2. Strategy for designing
universal overgo hybridization probes based on human-mouse
sequence alignments. Orthologous human and mouse genomic sequences are masked
for repetitive elements (indicated by X's) and then aligned. Regions with high
sequence conservation (indicated by vertical lines) are identified and used for
designing probes. When possible, a single 36-bp human sequence from each
alignment is chosen based on GC content and percent human-mouse sequence
identity. A subset of these sequences is then chosen to optimize for
inter-probe spacing (~30-40 kb). Three such conserved sequences are depicted in
the figure, with greater details provided (in the box) for the middle one. At
this stage, each selected 36-bp sequence is compared to all available human
genomic sequence to confirm that it is single copy. Overlapping pairs of oligonucleotide primers are then synthesized for each
sequence and used to generate double-stranded, radiolabeled
(indicated *'s) probes. The probes across a target region(s) are then pooled
and used to screen arrayed BAC libraries, allowing the isolation of individual
positive BACs. How the AUG_2003_b1 whole-genome probe set for
screening mammalian libraries was generated. Whole-genome human-mouse alignments (axtTight)
(Schwartz et al, Human-mouse alignments with BLASTZ. Genome Res 2003,
13:103-107 and Watertson et al, Initial sequencing
and comparative analysis of the mouse genome. Nature 2003, 420:520-562.) between the April 2003 assembly (UCSC version hg15) of the
human genome and the Feb. 2003 build (UCSC version mm3) assembly of the mouse
genome were downloaded from http://genome.ucsc.edu.
This alignment file was then modified to for use with a modified algorithm for
probe design, soop. Common repetitive sequences were
masked in the file and then, when possible, 1 candidate probe with >88%
human-mouse sequence identity from each ungapped
alignment was designed based on the human sequence. These candidate probes were
then compared to the April 2003 human genome build by megablast (megablast
-t 16 -N 2 -W 11 -e 0.6 -F F -D 3). The megablast
output was used to confirm the location of the probes in the human genome, and
tag each probe as 'unique' or 'non-unique'. Unique probes
had a single identical hit to the human genome assembly, no other hits with a bit
score above >40 and fewer than 5 hits with a score above 36. These are very
stringent criteria for calling a probe unique, and we feel that unique
probes, to the best of our knowledge, represent single-copy sequences in the
human genome and should be well suited for screening BAC libraries from other
mammals. Non-unique probes also had one identical hit to the human
assembly at the expected location, but had at least one other hit above 40 bits
or 5 hits above 36 bits. While the non-unique probes are not single-copy
in the human genome based on our criteria, we have kept them in our database
for potential use in regions of the human genome that are duplicated or for
isolating genes within a gene family. We do not recommend the use of non-unique
probes unless there is no alternative unique probe available. To
increase the number of unique probes, after masking just the non-unique
probe sequences in the human-mouse alignment file, candidate probes were then
designed only from alignments that yielded a non-unique probe.
Candidate probes were again compared to the human genome by megablast
and designated unique or non-unique. This recursive process was
repeated 2 times to yield 139,272 unique probes and 97,721 non-unique
probes. Figure 3.
Results of the experimental validation of a sample set of AUG_2003_b1 universal
probes. To test the efficiency of the
mammalian whole-genome probe set, AUG_2003_b1, n=48 probes were selected from
n=7 regions of the human genome for screening the marmoset (CHORI-259), galago (CHORI-256), rabbit (LBNL-1), bat (VMRC-7), shrew (SA_Ba), armadillo (VMRC-5), wallaby (ME_KBa),
and platypus (OA_Bb) BAC libraries. After primary and
secondary screens, probe-content information was merged with restriction-enzyme
fingerprint content maps. Based on this information, the success rate (the
fraction of probes tested that were positive for at least one BAC clone) in
each species was calculated, and is shown above. The distribution of the probes
percentage identity between human and mouse was slightly more enriched for
higher id probes than the content of the whole-genome set of unique probes (21%
versus 16% at 100% id, 19% versus 18% at 97% id, 19% versus 21% at 94% id, 23%
versus 22% at 91% id, and 19% versus 23% at 88% id). However, because optimal
physical spacing will greatly enhance the selected probes toward higher percent
identity, we believe this sample set reflects an accurate measurement of the
effectiveness of the unique whole-genome probe set. Representative clones have
been sent to the NIH
Intramural Sequencing
Center for sequencing to
confirm probe specificity.
Rationale
|
Table 2. Summary of Universal Probes in AUG_2003_b1 (Mammalian) |
|||
|
|
|
|
|
|
Human Chromosome |
Length w/o gaps (bp) |
Unique Probes |
Non-Unique Probes |
|
Chr1 |
218712898 |
12,156 |
9,401 |
|
Chr2 |
237043677 |
13,752 |
8,883 |
|
Chr3 |
193607233 |
9,624 |
6,298 |
|
Chr4 |
186580523 |
6,748 |
4,411 |
|
Chr5 |
177524972 |
9,125 |
5,989 |
|
Chr6 |
166880541 |
6,465 |
5,533 |
|
Chr7 |
154546299 |
6,675 |
4,770 |
|
Chr8 |
141694337 |
5,573 |
3,724 |
|
Chr9 |
115187719 |
6,151 |
4,501 |
|
Chr10 |
130710874 |
6,501 |
4,143 |
|
Chr11 |
130709420 |
7,735 |
5,451 |
|
Chr12 |
129328334 |
5,491 |
3,994 |
|
Chr13 |
95511656 |
3,512 |
2,462 |
|
Chr14 |
87191216 |
5,017 |
3,338 |
|
Chr15 |
81117055 |
4,943 |
3,929 |
|
Chr16 |
79890795 |
4,568 |
3,013 |
|
Chr17 |
77480855 |
5,655 |
4,433 |
|
Chr18 |
74534531 |
3,474 |
1,967 |
|
Chr19 |
55780860 |
1,970 |
1,596 |
|
Chr20 |
59424990 |
3,150 |
2,008 |
|
Chr21 |
33924747 |
813 |
713 |
|
Chr22 |
34352072 |
1,100 |
1,071 |
|
ChrX |
147686666 |
8,148 |
6,529 |
|
ChrY |
22761097 |
10 |
480 |
|
|
|
|
|
|
Total |
2832199938 bp |
138,356 |
98,637 |
OCT_2003_b2 Mammalian Whole-Genome Probe set.
A sample set of orthologous genomic sequences from human, mouse, rat, dog, cat, cow and pig were used to empirically optimize the universal probe design process using human-mouse-rat alignments. n=2863 36-bp probe sequences with n=7 or fewer mismatches between human and mouse, and for which rat, dog, cat, cow and pig sequences were also available were used as the basis of this process. For each substitution pattern between human-mouse-rat
(ie, human AAAAA
mouse AATTT
rat ATATC
pattern 12345)
a 'weight' was assigned based on the calculated percent identity for each pattern between the human nucleotide and the corresponding dog, cat, cow and pig nucleotide. The calculated values were:
sum of identical bases (dog, cat, cow,pig)/(total number of bases of Pattern# X 4)
Pattern 1=(67135+67461+66253+66623) /(72680X4)=0.9200
Pattern 2=(2472+2498+2399+2441) /(3003X4)=0.8167
Pattern 3=(1258+1263+1225+1215) /(1526X4)=0.8127
Pattern 4=(3476+3521+3399+3467) /(5885X4)=0.5889
Pattern 5=(318+330+304+310) /(501X4)=0.6297
A score was calculated for each probe by counting the patterns and then summing the corresponding 'weights'. Because rat and mouse are essentially equivalent distances from human, and only 0.004 separated the values for patterns 2 and 3, a single value, 0.8147 was used for both those patterns. The correlation of the probe scores and number of mismatches per probe in dog, cat, cow and pig was then calculated and compared to the correlation coefficient using a probe score based solely on the number of mismatches between human (probe) sequence and the mouse sequence. The correlation coefficient for the mouse mismatch score alone was n=0.5425073 and for the new matrix, n=0.564635, indicating that adding the rat sequence and using this matrix does provide a better basis for designing universal mammalian probes. While the increase in the correlation is not large, this basic scoring matrix strategy can be used with larger numbers and/or more informative combinations of species (such as human-mouse-dog).
The second major change to the probe design process was the selection of all probes that fell within the 0.44-0.56% GC range and met the minimum scoring requirement. In the previous build, only the 'best' probe was selected for each gap-free alignment between human and mouse. To provide the maximum number of probe options, we eliminated the 'best' criteria and now include all sequences that meet the set probe criteria.
This new algorithm was applied to the Multiz human-mouse-rat whole genome alignment (generated by W. Miller and J. Kent,(Blanchette et al. 2004. Aligning multiple geneomic sequences with the threaded block aligner. Genome Res 14:708-715. RGSPC. 2004. Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature 420:520-562) Human (UCSC hg15), Mouse (UCSC mm3), Rat (UCSC rn2), downloaded from the http://www.genome.ucsc.edu). Based on analysis of our test data set, a probe score cutoff value of 31.83 was determined to be more stringent than the 4 or fewer mismatches between human and mouse used in Aug_2003_b1 (ie, would include greater than 93% of all probes with 0,1,2,or 3 mismatches between human and mouse and exclude 60% of all probes with 4 mismatches between human and mouse). In addition, we have also edited the probe set to remove a small fraction of candidate sequences (<5%) that had properties that might compromise their general utility
(ie, a sequence that looked like this:GGCCGGGGGCCGCCCGGATATTATTTATAATATAT).
Specifically, probes with a gc_score (see soop.pl algorithm, OM40 (McPherson)) above 55.36.
Figure 4. Results of the experimental validation of a sample set of OCT_2003_b2 universal probes.
To test the efficiency of the mammalian whole-genome probe set, AUG_2003_b1, n=48 probes were selected from n=11 regions of the human genome for screening the marmoset (CHORI-259), galago (CHORI-256), rabbit (LBNL-1), bat (VMRC-7), shrew (SA_Ba), armadillo (VMRC-5), elephant (VMRC-15), wallaby (ME_KBa), and platypus (OA_Bb) BAC libraries. After primary and secondary screens, probe-content information was merged with restriction-enzyme fingerprint content maps. Based on this information, the success rate (the fraction of probes tested that were positive for at least one BAC clone) in each species was calculated, and is shown above. The test set of probes was selected to be an accurate sampling of the entire probe set (in terms of probe score). N Representative clones have been sent to the NIH Intramural Sequencing Center for sequencing to confirm probe specificity.
A numerical summary of this probe build is listed below.
|
Table 3. Summary of Universal Probes in OCT_2003_b2 (Mammalian) |
||||
|
|
|
|
|
|
|
Human Chromosome |
Length w/o gaps (bp) |
Unique Probes |
Non-Unique Probes |
|
|
Chr1 |
218712898 |
457,441 |
268,561 |
|
|
Chr2 |
237043677 |
420,924 |
222,490 |
|
|
Chr3 |
193607233 |
318,413 |
162,270 |
|
|
Chr4 |
186580523 |
198,668 |
103,338 |
|
|
Chr5 |
177524972 |
270,796 |
149,901 |
|
|
Chr6 |
166880541 |
213,399 |
139,525 |
|
|
Chr7 |
154546299 |
212,745 |
122,716 |
|
|
Chr8 |
141694337 |
179,965 |
91,364 |
|
|
Chr9 |
115187719 |
221,804 |
124,513 |
|
|
Chr10 |
130710874 |
208,993 |
105,613 |
|
|
Chr11 |
130709420 |
286,737 |
159,738 |
|
|
Chr12 |
129328334 |
211,643 |
124,001 |
|
|
Chr13 |
95511656 |
103,488 |
58,366 |
|
|
Chr14 |
87191216 |
178,789 |
100,094 |
|
|
Chr15 |
81117055 |
189,412 |
108,324 |
|
|
Chr16 |
79890795 |
159,073 |
91,888 |
|
|
Chr17 |
77480855 |
237,188 |
144,407 |
|
|
Chr18 |
74534531 |
99,938 |
48,999 |
|
|
Chr19 |
55780860 |
84,060 |
53,058 |
|
|
Chr20 |
59424990 |
112,647 |
58,712 |
|
|
Chr21 |
33924747 |
34,475 |
20,460 |
|
|
Chr22 |
34352072 |
61,558 |
39,617 |
|
|
ChrX |
147686666 |
224,952 |
149,438 |
|
|
ChrY |
22761097 |
112 |
8,855 |
|
|
|
|
|
|
|
|
Total |
2832199938 bp |
4,687,220 |
2,656,248 |
|
|
Table 4. Summary of probe scores for OCT_2003_b2. |
||
|
Score |
Unique Probes |
Non-Unique Probes |
|
31.83 |
18498 |
12059 |
|
31.84 |
55823 |
31366 |
|
31.85 |
47247 |
27331 |
|
31.86 |
243 |
144 |
|
31.87 |
3523 |
2334 |
|
31.88 |
4897 |
2868 |
|
31.89 |
4049 |
2336 |
|
31.91 |
211 |
159 |
|
31.92 |
344500 |
196685 |
|
31.93 |
41675 |
25794 |
|
31.95 |
1548 |
1296 |
|
31.96 |
68556 |
39478 |
|
31.97 |
7857 |
4694 |
|
31.99 |
227 |
214 |
|
32 |
5452 |
3309 |
|
32.01 |
330 |
223 |
|
32.02 |
436743 |
245136 |
|
32.04 |
88869 |
53423 |
|
32.05 |
4183 |
3044 |
|
32.06 |
73638 |
42661 |
|
32.07 |
301 |
165 |
|
32.08 |
15043 |
8745 |
|
32.09 |
423 |
326 |
|
32.1 |
4859 |
2865 |
|
32.12 |
621 |
383 |
|
32.13 |
351825 |
194616 |
|
32.14 |
177216 |
104856 |
|
32.16 |
10215 |
6712 |
|
32.17 |
45930 |
26272 |
|
32.18 |
25930 |
15562 |
|
32.2 |
1009 |
622 |
|
32.21 |
2586 |
1616 |
|
32.22 |
1071 |
670 |
|
32.25 |
300994 |
174414 |
|
32.26 |
23845 |
15111 |
|
32.28 |
761 |
490 |
|
32.29 |
38227 |
22927 |
|
32.3 |
2206 |
1559 |
|
32.33 |
1384 |
943 |
|
32.35 |
402291 |
230841 |