Genome browsers:
Entrez (NCBI)
UC-Santa Cruz (browser example)
Ensembl (Comparison of sequence updates)

Draft sequence of the publicly funded Human Genome Project
Nature 409, 860-921 (15 February 2001)

Compare with recent Mouse Genome sequence paper.

(Figures for Nature paper are direct links to Nature Website)

Draft genome sequence of physical map
Map covers 96% of genome, actual sequence spans 94%, including gaps
Eight times larger than all genomes sequenced before it

Used hierarchical shotgun sequencing

Genomic DNA was partially digested to form bacterial artificial chromosome libraries.
Random clones were mapped onto genome.
Specific BACs were chosen for sequencing based upon optimal overlapping of BACs on genome
Individual BACs were then sheared for shotgun sequencing and sampling.

This approach helps minimize some complications in sequence assembly of entire genome, due to repeats and large genome
Also avoids problems caused by heterozygous nature of human DNA source
BAC clones (100-200kb each) were generated at 20-fold coverage
Digested with restriction enzymes to generate a fingerprint for each BAC
Clones were aligned with each other based upon the fingerprint sequence
BAC clones were then sheared for subcloning and shotgun sequencing
Sequence was assembled for each BAC clone
Clones were then compared with their predicted fingerprints
The sequenced clones were aligned with overlapping sequence clones to construct a fingerprint clone contigs.
These in turn were aligned with the physical map of the chromosome by sequence matches to mapped STS sites

	2004 csome	contigs	N50 size
	1	37	56331646
	2	21	68373980
	3	5	100530261
	4	15	23438009
	5	6	41199371
	6	9	48945890
	7	12	47497097
	8	10	38454502
	9	42	39435726
	10	19	30112613
	11	8	38509590
	12	8	32815934
	13	5	67740325
	14	1	88290585
	15	11	53619965
	16	5	42003582
	17	10	21163833
	18	5	20074199
	19	4	31383029
	20	6	26144333
	21	4	28617429
	22	10	23276302
	X	21	27718692
	Y	17	6265435
	M	1	16571

Sequences were merged using Gigassembler program, which aligns sequence from one clone with adjacent ones

Quality

Sequence quality then described as N50, the length such that 50% of all nucleotides are contained in contigs of the least this size
Each nucleotide was assigned a quality score by the PHRAP sequence Assembly program.
91% of the draft genome sequence had a less than one in 10000 errors, based upon this score
96% had a less than one in 1000.
Still many gaps in the sequence, however most of smaller than a single BAC clone
Average gap size is believed to be about 25 kilobases
88 percent of known human cDNAs this could be aligned with a genome database

Analysis

Genome was analyzed for variations in G-C content
Variations were found in G-C content on a number of length scales and may represent mosaic origin of chromosomes
Dark Geimsa bands were found to be regions of low G-C content

Physical and genetic distance compared

It appears that longer chromosome have a lower recombination frequency, which causes all of chromosomes to have approximately one crossover per meiosis (figure)

Repeated DNAs - most transposon-derived

Long interspersed elements (LINES)
Transcribed from Pol II promoter, translated
Assembles core complex of endonuclease and reverse transcriptase on RNA
Then moves to nucleus to integrate into genome
Full size are 6-8 kb. There are 850,000 in the genome, representing 21% of the total genome length

Short interspersed Elements (SINES- e.g. Alu) shorter than LINE (100-300 bp)
Have Pol III promoter and 3'end similar to LINE so it can use LINE replication machinery
1.5 million copies in genome (13% of total length)

Retrovirus-like transposons (450,000 copies, 8% total DNA)

DNA transposons 300,000 copies (3% of genome)

Alu SINE elements appear more frequently in GC rich DNA

Can estimate ages of transposed elements based upon phylogenetic analysis.

Older Alus appear to have accumulated in GC-rich regions. (Selected for evolutionarily?)

LINES do not appear to follow same selection

Genes

Found genes for many non-coding RNAs:

497 tRNAs - mostly clustered
rRNAs not sequenced, because they occur as ~150 tandem repeats (repeats avoided in selection of BACs for sequence)
snRNAs and snoRNAs

Protein coding genes

Relatively easy for bacteria, where there are (almost) no introns
In that case, one looks for open reading frames (ATG-to-STOP) of a threshold size
ORFs can be evaluated for use of preferred codons and ribosome binding sites

Eukaryotes harder because of introns, fewer clues as to which ATG is right one (use 1st AUG of mRNA)
Yeast and C. elegans genomes fairly straightforward because introns fairly short
Known human genes have longer introns and more of them

First looked for known genes in RefSeq database (non-redundant database of human cDNAs)
Of 10,272 cDNAs, 9212 (92%) could be aligned, at least in part, with genomic sequence
85% could be aligned over the majority of their sequence
16% aligned in more than one location, possibly signs of pseudogenes or paralogs (genes duplications that may serve another function)

Got complete alignment of 5364 genes

Were able to calculate size and number of introns, exons, and UTR (untranslated regions of mRNA)

Peak exon size similar for human, fly and worm, but tail of distribution longer for fly and worm, so mean exon size is smaller for human. But conservation of peak size suggests optimal size of exon aids in processing mRNA.

Peak intron size is larger and more variable for humans (peak is 87, but mean is 3300bp)

Variation in intron size seems to be related to GC-content of genomic DNA
Exons sizes are constant throughout genome, but intron sizes are longer in AT-rich DNA regions

Splice sites

Were able to map splice sites by comparison to cDNA
Found 98% of 53,000 slice sites followed canonical GT-AG pattern for introns (GT at 5' end of intron, AG at 3' end)
0.76% use related GC-AG
only 0.1% use AT-AC, which requires variant snRNPs for splicing
Remaining 1% don't fit any particular rule and may be alignment or sequencing errors

Prediction of novel genes

Can try to identify coding regions and exons in genomic DNA using programs which utilize hidden Markov models (HMM), statistical models combing information on splice sites, codon bias, and lengths of introns and exons

Used Ensemble and Genie programs

Ensemble uses prediction of Genescan (an HMM program), then checks these predictions against ESTs, mRNAs and protein motifs in known databases
Merged these predictions (35,500) with those from Genie, which tries to match 5' end ESTs with 3' end ESTs to make full-length predictions
Also merged with known sequences from RefSeq

Came up with total of 31778 predicted
14,882 from known genes
12,839 from Ensemble
4,057 from Ensemble-Genie

Matches to mouse cDNAs which do not match known proteins may indicate proteins of novel function previously not identified

Comparison of protein coding capacity of sequenced genomes (proteomes)

There appear to be about twice as many proteins encoded by humans than fly or worm
Arabidopsis appears closer to humans in complexity
The difference in proteome size of humans vs flies and worms was surprisingly low
Estimates of number of human genes had ranged from 35,000 to 150,00 based on ESTs and other calculations. Conservative estimates appear to have been the most accurate.

Found distribution of proteins that could be traced to evolutionary ancestors by doing BLASTP search of all proteins against known sequences. Included scores with E< 0.001.

Found 233 genes that were similar to bacterial proteins, but not worm, fly or yeast.
Not due to DNA contamination, because not exact match to bacterial sequence (which were filtered out of sequence before assembly)
Confirmed presence of 35 by PCR
Found orthologs (proteins of apparent similar function in other organisms) in other vertebrates for many

So these appear to be cases of horizontal transfer of genes from bacteria to their vertebrate host.

Humans appear to have not so many more types of genes, but more members of similar gene families

Compared human gene locations to those for mouse known so far from mouse genome project

Found that some chromosomes were essentially same arrangement in rodents and humans. Others are greatly rearranged. Rodents are believed to have accelerated rearrangements in their genomes compared to most mammals

Mouse Genome sequence paper

Gives comparitive mammalian genome analysis.
Helps to identify genes and exons from human genome initially missed
Provides sequence information for mammalian model organism.

Some highlights:

Mouse genome 2.5 Gb vs human 2.9 Gb (more deletions in rodents)
Can identify regions of synteny between mouse and human for 90% of genome.
Both genomes have ~30,000 genes
99% of mouse genes have a human homolog (and vice versa)
80% of genes, the best match in human genome has that the same mouse gene as its best match (ie, they are orthologs)
Some genes appear to have evolved more quickly than random chance (immunity and reproduction).

Commercial Human Genome Project

Science v. 291, pp 1304-1351.

Celera company

Genome was sequenced using a shotgun approach

• DNA sequence obtained from random clones of sheared genomic DNAs
• The size of the genome (3 billion base pairs) makes assembling short stretches of sequence (500 to 600 base pairs) a computational challenge
• 27 million sequencing reactions were performed, yielding 15 billion base pairs of sequence
• Sequence was obtained from five individuals of different ethnic backgrounds

Key to assembling sequence:

• Pairs of sequencing reads obtained from opposite ends of cloned insert ("mate-pairs".
• Clones were obtained from three different size specific libraries: 2, 10, and 50 kilobases
• Knowing the distance between pairs of reads from individual clones helps in assembling contiguous sequences
• Raw sequence was trimmed of known vector sequence, E. coli sequence, and mitochondrial sequence
• Additional sequence data was obtained by shredding sequence from the public Human Genome Project

1. A whole genome assembly in which the entire sequence was assembled into the chromosomes
2. An alternative strategy sorted data based upon homology to know chromosomes, allowing assembly of individual chromosomes

A key hurdle in assembling sequence is avoiding misalignments caused by repeated sequences in the genome and chimeric clones.

Either of these problems can result in assembling connecting sequences from different regions in the genome together.
Repeated sequences were screened by removing known repeated sequences
Also, unknown repeats indicated by over representation of particular sequences
Unique sequences were then assembled and oriented using mate-pair information

Sequence scaffolds could be assembled based upon distances between pairs of reads
The initial scaffold covered 73% of the genome additional gaps were filled using reads which showed homology to scaffolds at one end
Additional gaps were filled with known sequence from the public database
Sequence scaffolds were mapped to chromosome by homology to known chromosome markers

Completeness was assessed by comparison to known finished sequences for chromosome 21 and chromosome 22 completeness was better than 95%
Accuracy will was determined by comparison to known chromosome 21 and chromosome 22 sequences and estimated at 99.96%

Gene prediction

GBCH723 Home Page