1. Sequencing the Maize (B73) Genome
Maize Genome Sequencing Consortium
Genome
Sequencing
Center

2. The Team WU Genome Sequencing Center (R. Wilson, PI)
Bob Fulton, Pat Minx, Sandy Clifton
Arizona Genome Institute (R. Wing)
Cold Spring Harbor Laboratory
D. Ware, L. Stein
R. McCombie, R. Martienssen
Iowa State University (P. Schnable & S. Aluru)
The Maize research community

3. The Plan

4. Progress as of 9/30/06

6. Agenda 9:00 – 9:15 Introductions and Project Overview (Rick Wilson)
9:15 – 10:15 Plans and Progress – WU/AGI/CSHL/ISU Project
Map and Tile Path Selection (Rod Wing)
Library Construction and Production (Lucinda Fulton)
Sequence Improvement (Bob Fulton, Dick McCombie, Rod Wing)
Data Submission (Joanne Nelson)
Annotation and Data Display (Doreen Ware)
Outreach (Rick Wilson)
10: :30 Break
10:30 – 11:00 Plans and Progress – DOE Project (Dan Rohksar)
11:00 – 11: Future Plans and Collaborations
Pat Schnable (by phone) - retrotransposons
11:30 – Noon Executive Session
Noon – 1:00 Working Lunch and Discussion
1: Depart for Airport

7. BAC-by-BAC Strategy to Sequence the Maize Genome
Maize B73 Genome (2300 Mb)
BAC library construction
(Hind III, EcoR I/MboI ; 27X deep ; 150kb avg. insert)
Genetic Anchoring in silico, overgo hybridization
Fingerprinting
~460,000 BACs
BAC End
Sequencing
~800,000
BAC physical maps (HICF & Agarose)
FPC databases
(Agarose and HICF)
STC database
Choose a seed BAC
Shotgun sequencing and finishing
STC database search, FP comparison
Determine minimum overlap BACs
Complete maize genome sequence

8. Map Summary Total Assembled Contigs: 721
Equal to 2,150 Mb, 93.5% coverage of 2300 Mb genome
Anchored: 421 ctgs, 86.1% the genome
average anchored contig size: 4.7 Mb
Unanchored: 300 ctgs, 7.4% coverage
average unanchored contig size: 0.56 Mb
189 of the 300 unanchored contigs are less
than 10 clones
Largest anchored contig 22.9Mb in Chr9
Largest unanchored contig 6.7 Mb
Total FPC Markers: 25,924
STS markers: 9,129
Overgo Markers: 14,877
Anchored markers: 1918

9. MTP Selection Seed BACs: 4000, done Mega Contig: 197, done
Clone Walking from Seed BACs: 2,800 done; in progress
Total clones picked = 6,997
On track to deliver 1000 clones/month until maze MTP is complete

10. Flowchart for MTP picking and Library Construction
Clone selection
(combine seed BAC and BAC end sequences with fingerprinting and trace files)
Clone picking (Resource Center)
GenBank BAC end sequence database
MTP sequencing
Seed BAC database
Library DNA production
Library DNA production
DNA shearing
Hfq sequencing
MTP BAC end database
Clone verification
Clone shipping
Continue shotgun library construction at WashU

11. Seed BAC Walking
In Agarose and HICF map, selecting large clones next to seed BAC
Blastn search of BAC end sequences against seed BAC sequences
Check blastn alignment for candidate clones
Check trace file for Dye blob
Check the Sulston score in HICF map for overlap
Check Agarose fingerprints to avoid overlap with large bands
Choose walking clone

12. Minimum Tile Path Pipeline
BAC End Sequence of potential BACs are BLASTed against the Seed BACs
Results are classified based on location on the FPC
A table for each BAC is created of filtered BLAST results with links to CMap and GBrowse
Blast results are imported into CMap and GBrowse with additional information such as trace files and FPCs

13. Minimum Tile Path Pipeline Usage
A table of alignments between the seed BAC and the BAC end sequences contains links to CMap and GBrowse.
CMap displays the FPC data for the seed BAC and the potential next BACs.
GBrowse provides an alignment of the BES with the seed sequence and displays the trace data.

14. Blast Results Table

15. Maize Production Sequencing
Shotgun of 19,000 BACs
Fosmid End Sequencing of 1 Million Reads
BAC End Sequencing of 220,000 clones

16. Maize BAC shotgun BAC DNA received from AGI or prepared at the GSC
Small Scale Library Construction
Production Sequencing - 1,536 reads/project
Automated Shotgun_done

18. To date 3,106 BAC clones are shotgun_done

19. Maize Fosmid Sequencing
Fosmid trays 0001 to 0471 were received from Messing lab
Initial QC was fine, but bulk shipment has failed to grow
Stamping results of the original trays show no growth
85 Fosmid ligations which represent ~250,000 clones were
received from the Messing lab, plating is underway
GSC Fosmid library construction has been completed and
represents 1M clones
Expected completion date is November of this year.

21. Maize BAC End Sequencing
BAC end sequencing will be completed next week
Total of 440,000 reads from two different libraries
Pass rate of 75% with an average read length 600 bases
Paired end read rate is ~70%

22. Sequence Improvement Pipeline
Shotgun_done triggers the prefinishing pipeline
Initial identification of “do finish” regions
Manual sorting and use of autoedit(Gordon) to break apart misassembly.
Autofinish(Gordon) used to choose directed reactions for all gaps and regions of low quality in “do finish” regions
Reassembly and 2nd iteration of prefinishing pipeline
Final identification of “do finish” regions and handoff to finishing pipeline

23. Clone Improvement through the Prefinishing Pipeline

25. Coverage (green)
Spanning Plasmids
End

26. EST sequence
GSS sequence
Do Finish
Repeat Tags

27. Alignment with cDNA read pairs
Alignment with End Sequences

28. Future Plans for Improved Throughput
Automated Shotgun-done status assigning
Overlap Evaluation at Prefinishing
Addition of Fosmid End Pairs at Prefinishing
Direct Sequencing for Unspanned Gaps
Additional Finishing Staff Hired at all 3 Centers

29. Maize clone submissions
clone status submission keywords
shotgun complete HTGS_PHASE1; HTGS_FULLTOP
2 rounds of prefinish HTGS_PHASE1; HTGS_PREFIN
in finishing HTGS_PHASE1; HTGS_ACTIVEFIN
finished HTGS_PHASE1; HTGS_IMPROVED
Query GenBank by keywords
zea mays[ORGN] AND HTGS_PREFIN[KYWD] AND WUGSC[CNTR]
zea mays[ORGN] AND HTGS_IMPROVED[KYWD] AND WUGSC[CNTR]
Restrict by date range:
zea mays[ORGN] AND WUGSC[CNTR] AND HTGS_FULLTOP[KYWD] AND 2006/09[PDAT]
zea mays[ORGN] AND WUGSC[CNTR] AND HTGS_FULLTOP[KYWD] AND 2006/09/26:2006/10/03[PDAT] |

30. HTGS_IMPROVED submissions
Pick a clonename, any clonename -
DEFINITION Zea mays chromosome 4 clone CH201-11H16; ZMMBBc0011H16
Center project name: Z_AF-11H16
Improved sequence is annotated on submission record
Where possible, contigs have been ordered and oriented based on read pairing. and these regions are designated as scaffolds.
Small contigs (<2kb) that don’t represent a clone end, don’t contain improved sequence, or are not part of a scaffold are removed from the final submission.
Contigs are screened for bacterial contamination

31. FEATURES Location/Qualifiers
source
/organism="Zea mays"
/mol_type="genomic DNA"
/db_xref="taxon:4577"
/chromosome="unknown"
/clone="CH C8; ZMMBBc0112C08"
misc_feature
/note="scaffold_name:Scaffold1"
misc_feature
/note="assembly_name:Contig245
clone_end:left
vector_side:T7"
gap
/estimated_length=unknown
misc_feature
/note="assembly_name:Contig240"
misc_feature
/note="scaffold_name:Scaffold2"
gap
misc_feature
/note="assembly_name:Contig250”
misc_feature
/note="Improved sequence."
misc_feature
misc_feature
/note="scaffold_name:Scaffold3"
.....

32. GenBank
1005 HTGS_FULLTOP
254 PREFIN_DONE
1532 ACTIVE_FIN
357 HTGS_IMPROVED

33. Ongoing work at CSHL BAC Annotations Levels Data Analysis Display
Project Management
Collaborations

34. BAC Data Analysis Ensembl Pipeline 3 inclusive phases of annotation
Level I: Display BAC information
Level II: Sequence-based annotations
Level III: Integrative annotations
Shiran Pasternak, Apurva Narechania, Joshua Stein

35. Application of Mathematical Repeat Analysis
Identifies novel repeats w/o dependence on curation.
Based on frequency of 20-mers in JGI WGS sequence
Correlates with presence of retroelements.
Can modulate threshold to optimize application.
AC148169
Statistics are based on FgenesH exons from High Confidence Gene Models (Aligning nearly end to end with a peptide in nraa) which are classified as TE or WH.
Apurva Narechania, Joshua Stein

36. Retroelement Annotation
Collaboration with Jeff Bennetzen and Philip SanMiguel
Classify retroelement families
Current list covers ~68% of genome
Ten most prevalent account for ~80% retroelement sequences
Ji, huck, opie, zeon, cinful, prem1, grande, xilon, gyma, giepum
Goal is to visualize the history of transpositions
Retro class are color coded.
LTR’s were identified and are shown in separate track (for clarity track is compacted and labels are turned off)
Numbers below glyph show the start/end position of the alignment relative to the coordinates of the retro sequence (expressed as a percent of length).
So if the retro sequence was 1000 nt long then 0-15 would represent position 1 to 150 of that sequence and would represent position 450 to 630 of the sequence.
This allows us to reconstruct the giepum element as once being a full length and intact element into which a full-length ji element and opie elements have inserted.
Statistics on genome coverage are based on pilot 100 BACs
Giepum element interrupted by ji and opie in AC148166
Joshua Stein

37. Whole Genome Alignments
Wobble Aware Bulk Aligner (WABA)*
TIGR Transcripts Rice
WABA alignments Maize
Distinguishes between:
low similarity regions (grey)
high-similarity regions (medium blue)
high similarity regions w/ wobble-base mismatch of coding regions (green)
*Kent, WJ & Zahler, A.M. (2000). Genome Res. 10:
Joshua Stein

38. Whole Genome Alignments
BLASTZ* with AXTCHAIN** & CHAINNET**
Sensitive gapped BLAST algorithm designed for aligning long sequences.
Accommodates long gaps & overlapping gaps, inversions, translocations, & duplications
*Schwartz, S et al. (2003). Genome Res. 13:103-7
**Kent, WJ, et al. (2003). PNAS 100:
Example of BLASTZ(net) display in Ensembl.
Cg is chicken and Cf is dog and the reference genome is human.  I chose these because the primates were so close to human that essentially you got a solid pink line across the screen (not very interesting).  

39.
Sequenced BAC
FPC Contig
Virtual Bin
Core Bin Marker
Chromosome
Synteny Views
Main Navigation bar is accessible from every page
Contains multiple entry points to the genome

40. MapView
Displays statistics by chromosome and provides entry points based on a single chromosome

41. CytoView
Provides detail information on features anchored to the FPC map.
The side bar highlights the location on the chromosome and provides page specific functionality including data export.
The Detailed view is customizable, tracks can be added or removed by the users.
Feature contain drop down menus that contain general information as well as provided internal links, and external links.

42. ContigView
This view is based BAC coordinated and displays annotation levels II and III.
The header contains the Clone name in the physical map, GenBank Accession, and Chromosome and FPC contig information.
Detailed view offers semantic zooming, customizable and provides links to other views and information resources.

43. SyntenyView

44. Upcoming Features Notification System Release Users are notified
October 2006
BlastView
December 2006
BAC Annotation
Level II January, 2007
Level III annotation April, 2007
WG alignments June, 2007
BioMart
January, 2007
NSF collaborations
TwinScan annotations: March, 2007
Maize Optical Map: July, 2007
Full-length cDNAs: December, 2007
Notification System
Users are notified
When a region of interest is updated
When markers are aligned to a specific sequence
January, 2007

45. Hardware Environments
Software
Developed locally
Managed with source control
Frequent releases to staging environment
Quarterly production releases
Data
Timed analysis on staging environment
Mirrored weekly on production
The two maize machines (cannon and ascutney) are Dell Poweredge 2850's (2U rackmount). The have:
- 8GB of memory
- 1.5 TB of redundant disk space
- Dual 3.8Ghz Intel 64bit processors
The machines are backed up utilizing Atempo's Time Navigator backup software. The backups are performed over a gigabit switched network and the data is written to SDLT tape housed in an Overland Storage tape library. Full backups are performed every four weeks and incrementals three times per week.
42 IBM blade servers each containing:
- 2 dual core AMD 2.0 Ghz 64 bit CPU's (84 CPU's. 168 cores total)
- 4GB memory
- 73GB SCSI disk drive
Doreen has 3 BC, each with 14 blades. Each blade contains: two 2.0GHz Dual core Opteron 64's, with 4 x 1GB memory and a single 73Gb SFF SCSI disk. These are integrated into the HPCC. Overall there are 33+3 BC's in the HPCC all quad GigE connected to a central Cisco 6509 core. There are 2 management nodes and 2 storage nodes. The storage nodes are 32 bay Aberdeen units filled with 500GB SATA II drives in RAID6 configurations. How best to use the space, how any backups will be done, etc. is yet to be determined -- depends on user requests and usage; but overall there will be ~20TB available.
Shiran Pasternak, Apurva Narechania

46. Quality Assurance Unit-testing framework Software Quality Control
Binary assertions
Failure report and automatic notification
Software Quality Control
e.g., code retrieves correct data from the database
Data Quality Control
e.g., clone in Genbank record exists in FPC map
Shiran Pasternak

47. Project Management Mantis Bug Tracker
Manage tasks using priorities, severities, and resource allocations
Automated submission of issues using feedback form
Generation of progress reports

48. Project Management Wiki Enhances group communication
Meeting notes, flowcharts, specification documents
Maintains history of specifications and design decisions
Seamless editing

49. Collaborations
MaizeGDB (Iowa State University, University of Missouri)
C. Lawrence
Maize Optical Map (University of Wisconsin)
D. Schwartz
Maize Transposon Annotation (University of Georgia, Purdue)
J. Bennetzen, P. San Miguel
Ensembl (EBI)
E. Birney
Vmatch for Mathematical Repeats (University of Hamburg)
S. Kurtz
Maize Full Length cDNA project (Arizona Genomics Institute)
Y. Yu
TwinScan (Danforth Plant Science Center)
B. Barbazuk