1. Public data and tool repositories Section 2 Genome Browsers

2. Problems from last section
Query Entrez Gene with the following two queries separately and then explain the differences between the two results using a logical NOT operation:
tyrosine kinase[Gene Ontology] AND human[Organism]
cd00192[Domain Name] AND human[Organism]
Retrieve the APP gene record from NCBI and use the Display dropdown menu to display Conserved Domain Links. Use the ids of the listed domains to query Entrez Gene for records with the same domains.
Use the SNP Geneview link at NCBI to identify coding SNPs in the APP gene. Which SNP is missing from this display which was present in the Ensembl APP protein record?
Use the Homologene link at NCBI to identify possible functional orthologs for human APP. How does this list compare to the Ensembl list of orthologs that we reviewed previously?

3. Review of last section example: human APP gene
NCBI Entrez databases
Constructing queries
Gene, Nucleotide and Protein
RefSeq
EBI/Ensembl
Finding genes
Viewing Genes, Transcripts, Exons, Proteins and SNPs
Common id and data formats

4. This section Genome assembly and genome browsers
Promoter/enhancer analysis example
More information

5. Genome Build Process
Organism sequence data is assembled into contiguous pieces (contigs)
Contigs are mapped to genomic features and the coordinate system is assigned
Unmapped sequence data be assigned to artificial chromosomes
Assembly is improved as more sequence data is available
Entrez Genome Project

6. Genome Browsers
Make millions of sequences available through easily accessible, user-friendly interfaces
Provide genomic sequence, exon structure, mRNA sequence, EST and SNP data via web-based text search interfaces
Options available for local installs

7. Commonly Used Browsers
The Entrez Map Viewer
The EBI/Ensembl browser
The UCSC genome browser

8. NCBI Map Viewer
Integrates feature identity information with whole genome view
Allows one to view and search an organism's complete genome
Displays chromosome maps
User can zoom into progressively greater levels of detail, down to the sequence data for a region of interest.
Focus more on individual sequences
Worked example: Querying for APP gene in NCBI Map Viewer
1. Go to
2. Select Homo sapiens in Search drop down menu
3. Type APP in the for: text box
Select Go button
Map Viewer will display all hits to “app” on all chromosomes in all human assemblies
Number of hits per chromosome is displayed in red under the chromosome number, the location of each hit is highlighted on the chromosome and the hits are listed for each assembly
Type 21 into on chromosome box
Select Find button
Select the APP Map Element of type GENE
The map viewer now displays the chromosome view with APP highlighted.
Things to note:
a. NCBI displays its data along the vertical axis (unlike UCSC and ENSEMBL)
Each feature has a list of links to its right, these are link outs to detailed feature info, and will vary based on the feature type.
12. The list of links for APP include:
a. APP – this links to the feature specific page, in this case the gene entry.
b. OMIM - online mendelian inheritance in man: This database is a catalog of human genes and genetic disorders and is curated manually.
c. HGNC- HUGO gene record. HUGO is the authority that issues canonical gene names. This is fairly recent, so many genes have old names and pseudonyms... these alternate names should be listed in HUGO. (ex AD1)
d. sv - detailed nucleotide record for feature. Sort of analogous to the UCSC view.... but just for a single feature.
e. pr - NCBI protein record. Genbank file for the protein record.
f. dl - download feature info. This page allows you to download the feature(s) in the range of interest. You can specify formats and other settings.
g. ev- Evidence Viewer. Supporting data for this feature. mrna seq, protein seqs, EST.
h. mm - model maker view: Model Maker allows you to view the evidence that was used to build a gene model on assembled genomic sequence, and to create your own version of the model by selecting exons of interest.
i. hm - homolog gene. Lists related genes in all species
j. sts – uniSTS
k. CCDS - CCDS entries are manually curated gene sets that are considered to be correct. Each splice variant in the CCDS is experimentally validated, and no genes with ambiguous translation will make it into this set. It covers ~ 3/4 of the known Hs genes to date
l. SNP - dbSNP records in this range
Ex: Looking at the APP gene in the NCBI Map Viewer

9. EBI/Ensembl Browser
Provides access to sequence data from ~40 organisms
Includes the human genome sequence and data from all the commonly used experimental organisms
Displays the location of genes, variations and other sequence features within genomes
Greatest strengths:
browsing of large genomic contigs
comparative genomic features
Worked example: Querying for APP gene in EBI/Ensembl browser
Go to
Select Homo sapiens in drop down menu
3. Type APP in the text box
4. Select Go button
5. Choose the 1st record
6. This page displays the full info for APP: transcript, genomic info, cross links, etc.
7. In the Genomic Location field(5th field down in the report), choose the 1st link.
8. The genome browser pops up centered on APP.
9. Notice this browser has a horizontal orientation, and has 4 sections:
Chromosome, Overview, Detailed, and Basepair
Expand the base pair view
Explore the ‘Detailed’ panel:
a. Allows you to set the display information to view
b. Use the drop down menus at the top of the box to choose display fields
c. Choose to display SNPs
c. Display setting can be set with these menus
d. Finally, you may also export the image
12. Explore the interactive display
13. The right hand columns offer links to several more tools: comparative alignments, data export utilities, and link outs to other browsers
Ex: Looking at the APP gene in the EBI/Ensembl Browser

10. UCSC Genome Browser
Strength is genome position-based data aggregation:
Data positioned on “best” genome build and organised into “tracks”
Outside data tracks
Genome builds
Genes, known and predicted
mRNA
Expression and regulation
Variations and repeats
Inside data tracks
Known Genes
Comparative genomics
Custom tracks
Worked example: APP in the UCSC Genome Browser
1. Go to
2. Click on the Genome Browser link on the left sidebar
3. Verify that genome: Human and the latest assembly (currently, Mar. 2006) are
selected
4. Type in app into the query box and click the Submit button
5. In the list of results, click on the link for APP, isoform a (but isoforms b and c are just
as good since all the positions are the same)
6. Orient the participants to the display…
7. at the top are some browsing and zooming controls
8. on the next line are the position jumping controls
9. next is the chromosome overview
10. then we have the browser window with data tracks displayed as position-registered,
layered horizontal lines
11. and at the bottom are track display controls.
12. Let’s make sure we’re all looking at the same tracks…
13. Scroll down to the track display controls and make sure everything is set to “hide” by
clicking on the hide all button
14. Set the following tracks to “dense”: Base Position, Known Genes, RefSeq Genes,
Ensembl Genes, Human mRNAs, Spliced ESTs, Human ESTs, UniGene, TFBS
Conserved, Conservation, Most Conserved, Repeat Masker, Simple Repeats
15. Click on the refresh button
16. Point out the different marks on the gene tracks: exons: large ticks, UTR: small ticks,
line: intron, arrow heads: transcription direction
17. Click on the Known Genes, RefSeq Genes and Ensembl Genes to expand them to
packed display
18. Known genes are UCSC pick of “best” gene models
19. Ensembl gene predictions are analogous to RefSeq, but they are made by EMBL
(more about that later)
20. Human mRNAs, Spliced ESTs and Human ESTs are derived from sequencing
records.
21. Notice that the UniGene clusters for this sequence are fractured and don’t combine
exons. Possible explanations include too few mRNA and EST sequences to form
robust clusters, bad EST coverage of exons won’t allow good sequence overlaps, or
misalignment of cluster sequences with the genome.
22. To check out the number of EST sequences, click on Human ESTs. It appears that
there is very good EST coverage of this gene, but that they are very fragmented with
numerous 3’ ends. This could confuse the UniGene clustering algorithm since it
weights 3’ ends more highly than 5’ ends.
23. To check out how big the exons are, click on one of the RefSeq sequences, and then
click on the linked Entrez Gene id.
24. In the new browser window, change the display to Gene Table and scan through the
exon lengths. No problem here, since the exons seem to be about average size (about
100nt long).
25. To check for misalignment of cluster sequences with the genome, take note of one of
the RefSeq sequence ids (say NM_000484), click on one of the UniGene ids, and then
the Human UniGene id.
26. In the UniGene record browser in the new browser window, search for the RefSeq
sequence. We find that Hs is large cluster that defines APP.
27. Close that browser window and search for Hs in the UCSC genome browser.
It is not found and therefore the problem must be misalignment.
28. Press the browser’s back button until you’re showing the APP gene again
29. Click the hide all button
30. Set the following tracks to “dense”: Base Position, Known Genes, RefSeq Genes,
Ensembl Genes, TFBS Conserved, Conservation, Most Conserved, SNPs, Repeat
Masker, Simple Repeats
31. Most tracks have the following display types: hide, dense, squish, pack, full
32. The TFBS Conserved track is a good one to try out the different track display types--
have participants try each type on their own--remember, nothing will happen unless
you click the refresh button
33. Make all the tracks dense again, we’re going to look upstream for possible regulatory
regions.
34. Click on the TFBS Conserved and Conservation tracks to flip them into pack display
35. The Conservation track shows a histogram with degree of conservation after
comparison to other vertebrate genomes.
36. The TFBS Conserved track shows high scoring transcription factor binding sites in
regions of high vertebrate conservation. These are computationally derived.
37. Click on the double arrow right to pan the browser view to the right.
38. Identify the TFBS Conserved site upstream of the putative transcription start site for
APP.
39. This is V$MEF2_01. Click on it. This is the record for myogenic enhancing factor 2
family of transcription factors.
40. To see if this is known to be related to APP, open a new browser window, go to
NCBI and search all databases for MEF2 AND APP.
41. You get two hits in PubMed and two hits in Gene. Click on the PubMed hits.
42. Click on the Burton et al paper to read the abstract and see that APP seems to mediate
phosphorylation of MEF2, which regulates neuronal survival. There was differential
activity of mutated APP in familial AD. There is also a GeneRIF record in the APP
gene record which references this paper.
43. Hypothesis: perhaps APP activates MEF2 and MEF2 negatively regulates APP
transcription?
Ex: Looking at the APP gene in the UCSC Genome Browser

11. APP Upstream Region
15kb
Worked example: Extracting and aligning human and mouse APP upstream regions
1. First we will extract DNA sequence upstream of human and mouse APP genes
2. Go to NCBI at
3. Query for APP[Gene Name] AND (human[Organism] OR mouse[Organism])
4. Click on the Gene results
5. Click on the human APP gene record
6. Click on the gene’s chromosome sequence, and select FASTA format
7. Note that we are viewing the sequence on the minus strand of DNA
8. We will extract sequence 10kb upstream of the transcription start site (TSS)
9. Copy the coordinate in the “Range to” text box, which is our working TSS
10. Paste it into the “Range from” text box
11. Add to the number in the “Range to” box (remember we on the minus strand)
12. Click the Refresh button
13. Select File in the “Send to” dropdown menu
14. Confirm that there is a file called “sequences.fasta” on the desktop
15. Click the browser’s go back button until you are at the Gene database results
16. Click on the mouse APP gene record, and repeat steps 5-12.
17. Confirm that there is a second file on the desktop called “sequences(2).fasta”
18. Now we will confirm that we extracted the correct sequences
19. Open a new browser window and go to UCSC Genome Browser at
20. Click on the BLAT link on the left sidebar
21. Make sure the genome dropdown menu is set to human
22. Click the Browse… button below the large text area and select the human
“sequences.fasta” file on the desktop
23. Click the submit button
24. Click on the browser link on the first hit in the list of results
25. Your query sequence is at the top of the genome browser viewer as the black box and
arrows indicate direction of your sequence
26. Click on the zoom out 3X button and verify that your sequence is 10kb in length and
directly upstream of the APP TSS
27. Click the browser’s back button until you are at the BLAT Search Genome page
28. Select the mouse genome and browse for the mouse “sequences(2).fasta”, and then
repeat steps 23-26
29. Now we will do an alignment using some comparative genomics tools
30. Go to DCODE.org at
31. Click on the zPicture link on the left sidebar
32. We will enter the human DNA as sequence 1 and mouse as sequence 2
33. In the Sequence 1 box, click on the Browse… button, go to the Desktop and select
“sequences.fasta”
34. In the Sequence 2 box, do the same but select “sequences(2).fasta”
35. Click the SUBMIT button at the bottom of the page, and you’ll get a queue message
36. Click on the “click here” to refresh link, and you’ll eventually see a results page
37. Click on the Dynamic visualization icon
38. The graphic on this page shows a comparison between our human and mouse DNA
sequences as percent identity and red shaded segments are segments showing high
identity.
39. Click on one of the red-shaded sections and you’ll retrieve a regional alignment and
the sequence segments.
40. Click on the browser’s go back button
41. Click on the Dot-plot icon
42. A dot-plot is the traditional way to view pairwise sequence comparisons, particularly
whole genome-genome, since it is excellent for viewing duplications and inversions.
One sequence is on the horizontal axis and the other on the vertical axis. A dot
indicates a high degree of sequence identity.
43. Click on the browser’s go back button
44. Click on the rVista icon
45. Make sure that TRANSFAC professional library is selected and click the SUBMIT
button
46. Since we are interested in the MEF2 transcription factor, scroll down to the M section
and check the box next to MEF2 and click the SUBMIT button
47. Click on the CHECK button
48. Click on the Dynamic Visualization icon
49. Notice that this is a similar display to what we have already seen, except there is an
extra track above the image
50. Check smooth plot in the Picture section
51. Check conserved and all in the Show section
52. Check flip above the SUBMIT button so that we see the sequence as we extracted it
from NCBI (zPicture flipped our sequence to the positive strand for some reason)
53. Click SUBMIT
54. The ALL track shows all MEF2 sites detected in the human sequence, the
CONSERVED track shows MEF2 sites that were found in both sequences in a
conserved region
55. Click on the browser’s go back button
56. Click on the linked Highlight in the Alignment section
57. Click the SUBMIT button
58. Scroll down the alignment of the conserved segments and note the single conserved
MEF2 site in blue.
Ex: Extracting and aligning human and mouse APP upstream regions

12. Promoter/enhancer analysis approaches
Same gene, multiple species
Assumed evolutionary conservation of non-coding regions
Can use pairwise or multiple alignment method
Examples:
Precomputed: UCSC conservation tracks
Dynamic: eg, rVista
Different genes, same species
Typical output as co-expressed clusters from microarray data
Looking for over-represented, small binding sites
Much better results if looking for a pattern or clustering of multiple sites
Motif-finding algorithm, eg, MEME

13. Tutorials NCBI EBI UCSC Bulk Downloads Field Guide
Information and tutorials
Science Primer
EBI
2Can Tutorials
UCSC
Genome Browser User’s Guide
Bulk Downloads
Bulk Downloads Tutorial

14. IN CLASS EXERCISE
1. Do all three browsers show the same number of transcript variants for: APP, EGFR, TP53?
2. How many SNPs appear in the 5’ UTR of APP?
3. What is the lowest conservation score in APP exon 2?
1) Do all three browsers show the same number of transcript variants for: APP, EGFR, TP53?
i) Open the UCSC browser
ii) Search for Gene name
iii) Open link for isoform A of gene.
iv) View UCSC known gene track
v) Count the number of UCSC known genes
vi) Click ENSEMBL link (open in new window)
vii) Click on the gene name in the detail view.
(a) Open gene record.
(b) This page lists the number of transcript id's.
viii) Click NCBI link (open in new window)
ix) The map viewer should be open to the gene record.
x) Click gene name. This page will list the transcripts.
xi) Repeat Steps ii through
TP53 EGFR APPUCSC Ensembl NCBI
2) How many SNPs appear in the 5’ UTR of APP?
i) Open UCSC browser
ii) Search for APP
iii) Choose 1st record
iv) Turn on the SNP trakc to full
v) Refresh the display
vi) Zoom in on RIGHT end of gene (this gene is in the - strand, so the 5' UTR is on the right in the genome view!).
vii) Count the snps
3) What is the lowest conservation score in APP exon 2?
i) Same as in ex 2, but zoom in to 2nd exon from RIGHT end Turn on the conservation track (multiz align)
ii) Refresh the display
iii) Zoom into the lowest point of the conservation graph until you are at the single base level view
iv) Click on the conservation track at this pointThis opens the conservation track info page. There should be a link at the top for 'conservation score
v) Follow link and get the conservation score from the table.