Transcript WGA vs. CSA - Western Washington University

Compartmentalized Shotgun Assembly
?
?
?
CSA
Two stated motivations?
?
Matcher matched…
…matched Celera reads with PFP BACTIGS,
– 20.76 million Celera reads matched (76%),
– 0.62 million had a mate pair that matched,
• 2.97 million Celera reads were unique and unscreened,
– 1.189 Gbp of unique DNA sequence, at 5.11X yields a
predicted 240 Mbp of unique Celera sequence.
Combining Assembler assembles…
“…Celera and PFP sequence for a transient assembly”
…first, Celera reads,
– are checked for over-collapsed regions,
• sequences with Mate Pairs that match region are kept,
• more mate pair matches = higher value assembly,
…then Celera reads are combined with PFP reads,
• “Greedy” program recognizes highest value assemblies first in order
to build contigged sequence,
…then “Stones” to fill the gaps.
Results…
PFP vs. CSA
• The GenBank (PFP) data for the Phase 1 and 2
BACs yielded an average of 19.8 bactigs per
BAC, of average size 8099 bp,
• Application of the Combining Assembler resulted
in individual Celera/BAC assemblies being put
together into an average of 1.83 scaffolds (median
of 1 scaffold) per BAC region consisting of an
average of 8.57 contigs of average size 18,973 bp.
pp. 1313, 1st column, last paragraph
Compartmentalized Shotgun Assembly
?
Celera Unique Scaffolds
WGA
• The 5.89 million Celera fragments not matching the
GenBank data were assembled with the whole-genome
assembler.
• The Celera assembly resulted in a set of scaffolds totaling
442 Mbp in span and consisting of 326 Mbp of sequence.
More than 20% of the scaffolds were >5 kbp long, and
these averaged 63% sequence and 27% gaps with a total of
302 Mbp of sequence.
Compartmentalized Shotgun Assembly
?
?
Tiler tiles…
• Scaffolds into larger components using
– Mate End Pairs,
– BAC-end pairs,
– STS,
• Heuristic: a rule of thumb, simplification, or educated
guess that
reduces or limits the search for solutions in domains that are difficult
and poorly understood. Unlike algorithms, heuristics do not guarantee
optimal (or even feasible) solutions and are often used with no
theoretical guarantee.
Compartmentalized Shotgun Assembly
•3,845 Components
• shredded, WGA
*
93%
• > 100 kbp Scaffolds;
– 92% sequence, 8% gaps,
– > 49% gaps < 500 bp,
– 105,264 gaps, 1,935 scaffolds,
– > 62% gaps < 1 kb,
– 1.3 Mbp scaffold size, 23,242 bp – No gap larger than 100 kbp.
contig size.
How do you compare assemblies, and why?
WGA vs. CSA
• This gives some measure of consistent coverage:
– 1.982 Gbp (95.00%) of the WGA is covered by the CSA,
– 2.169 Gbp (87.69%) of the CSA is covered by the WGA.
• Only 31 scaffolds were ~unique to an assembly,
• 295 kb (0.012%) CSA inconsistent with WGA,
• 2.108 Mb (0.11% WGA inconsistent with CSA,
small
regions
Overall, CSA slightly better than WGA…
Why?
How does the CSA compare with the Clone-by-Clone approach?
Hierarchical Clone-by-Clone
Map First: then sequence
Whole Genome Assembly
Sequence First: then map
Mapping Scaffolder
GM99 and fingerprint maps
Mapping Scaffolder
GM99 and fingerprint maps
?
Tab. 4
Assembly and Validation Analysis
…did it really work?
• Completeness: % of euchromatic sequence in the
assembly,
– estimate the size and # of gaps (Table 3),
WGA
CSA
PFP
91 % Sequence
92.2 % Sequence
92.5 % Sequence
9 % Gaps
7.8 % Gaps
12.9 % Gaps
102,068 Gaps
116,442 Gaps
Small gaps (554 bp) = 145,514 Gaps,
Large gaps (35 kb) = 4076 Gaps.
Assembly and Validation Analysis
…did it really work?
• Completeness: % of euchromatic sequence in the
assembly,
– estimate the size and # of gaps (Table 3),
– compare to “finished” sequences of 21,22
• 3.4 Mb gaps, 75% gaps are repeats,
– match with STS data (ePCR, BLAST),
• 93.4% tested found assembled, 5.5% in “chaff” = 98.9%,
• Correctness:
– Mate-Pair analysis.
Mate Pair Analysis
Valid: correct orientation and correct distance + 3 SD
2.7% were found to be invalid.
CSA vs. PFP
What does this show?
Chromosome 21
Blue: Gaps
Violations:
Red : misoriented
Yellow: distance
PFP
CSA
Green: Same Order,
Orientation
Yellow: Same
Orientation
Red: Out of Order,
Orientation
Chromosome 8
PFP
CSA
PFP
CSA
What’s the take
home message?
PFP
CSA
Blue: breaks
Red: gaps > 10kb
Fig. 7, key
Fig. 7
Gene Prediction and Annotation
Why’s it So Hard to Find Genes?
• Exons/Introns,
• Alternative Splicing/Termination,
• Alternate transcription start/stop sites,
• Tandem Repeats, Psuedogenes, etc.
• We don’t really understand all there is to know
about gene and genome structure,
• etc.
Gene Number Predictions?
…before PFP, WGA or CSA
Textbooks: ~100,000
Upgraded to 142,634? EST data
“…counts [that] fall far short…”
EST Data --> 35,000
35,000 genes based on the density
of Chromosome 22
28, 000 - 34,000 Humans vs.
pufferfish
Automated Gene Annotation
OTTO
Tell me how it works.
How was it validated, including Table 7.
…if necessary, use the Online Primer and other
NCBI resources to broaden your understanding,
– cDNAs, ESTs, RefSeq, Protein Sequence Databases,
BLAST, etc. are described in appropriate detail on the
WEB.
Questions?
Repeat Resolver
...most of the remaining gaps were due to repeats.
“Rocks”
Use “low Discriminator Value” contig sets to fill gaps,
- find two or more mate pairs with unambiguous matches in the
scaffold near the gap (2 kb, 10kb or 50 kb), (1 in 107),
“Stones”
- find mate pair matches 2 kb, 10 kb, and 50 kb from gap, place the
mate in the gap, check to see if it’s consistent with other “placed”
sequences.
Repeat Resolver
...most of the remaining gaps were due to repeats.
“Rocks”
Use “low Discriminator Value” contig sets to fill gaps,
- find two or more mate pairs with unambiguous matches in the
scaffold near the gap (2 kb, 10kb or 50 kb), (1 in 107),
“Stones”
- find mate pair matches 2 kb, 10 kb, and 50 kb from gap, place the
mate in the gap, check to see if it’s consistent with other “placed”
sequences.