Introduction

There are principally three types of biological sequences, with the information flowing as outlined in the Central Dogma of molecular biology:

Once a new sequence has been determined, there are various ways of trying to find the function:

• Through a comparison of how well it matches another sequence of known function (alignment methods).
• By looking for characteristic patterns within the sequence (de novo methods).
• Prediction of the sequence structure (ab initio methods).

In this talk (as well as the one after the break), I will focus on only the latter two approaches - that is, looking for patterns in DNA sequences and predicting local DNA structures based on the DNA sequence. Both talks will be about information in the DNA sequences within the context of sequenced bacterial chromosomes. Note that this is only one tiny fraction of the much larger subject area of bioinformatics.

Genomics is simply the study of complete genomes.  More recently, genomics has meant the analysis of the complete DNA sequence of an organism's genome.  This field wasn't really possible until the publication of the genome of Haemophilus influenzae in 1995.

A Few Words on the speed of DNA sequencing

I know this is a bit of a digression from MICROBIAL genomes, but I want to try and add a bit of historical perspective. In 1977, Fred Sanger sequenced the first bacteriophage (phiX174, 5386 bp long), for which he later won the Nobel prize.  Although this was a dramatic improvement over the conventional methods, this was still very slow, compared to the amount of information in a single human cell.

About a decade later, the human genome project was launched; this was an international effort, and the U.S. would pay about $200,000,000 per year for 20 years!  Most of this investment was in technology to speed sequencing, which in fact has been realised.  Within a few years, it is likely that it will be possible to read the entire DNA sequence of a human cell, in a few hours.

A Timeline of The Human Genome Sequencing Project

YEAR	# human genes mapped to a definite chromosome location	# years it would take to sequence the human genome
1967	none	sequencing not possible yet
1977	3 genes mapped	4,000,000 years to finish at 1977 rate
1987	12 genes mapped	1000 years to finish at 1987 rate
1997	30,000 genes mapped	50 years to finish at 1997 rate
2001	~45,000 genes mapped	Finished. (kind of) Two versions: Celera and "Public". Agilent Technologies announces that they are developing "nanopore technology", which could allow the entire human genome to be sequenced in a few hours!

A look at genome sequencing since 1994 (including bacteria, archaea, and eukaryotes):

YEAR	# GENOMES Sequenced	Running Total
1994	0	0
1995	2	2
1996	2	4
1997	5	9
1998	8	17
1999	13	30
2000	23	53
2001	42	95
2002	>100	>200

Currently (4 September, 2002) NCBI lists 166 BACTERIAL genomes in its database.

Note that I've only listed the PUBLICLY AVALIABLE genomes. There are probably more than a THOUSAND bacterial genomes which have been sequenced by various companies which will never make it into the public domain.

What is missing?

There are (at least) TWO things missing: genomes from ecologically abundant and diverse niches, and larger genomes.

On the size variation within bacterial genomes

The size of most bacterial genomes varies considerably. For example, consider the four Escherichia coli genomes which have been sequenced (so far!), as shown in the table below.

Organism	%AT	Size (bp)	Atlas	Number of genes	Coding density	Reference
Escherichia coli Strain: K-12, isolate W3110 DDBJ     NCBI tax	49	4,636,552	Genome Atlas	4085	79%  1135 bp/gene	-
Escherichia coli Strain: K-12, isolate MG1655 U. Wisconsin     TIGR cmr     NCBI tax     NCBI entrez	49	4,639,221	Genome Atlas	4397	87%  1055 bp/gene	Science 277:1453-1474 September, 1997 [PubMed]
Escherichia coli Strain: O157:H7 (substrain EDL93) U. Wisconsin     NCBI tax     NCBI entrez	49	5,529,376	Genome Atlas	5283	86%  1047 bp/gene	Nature 409:529-533 January, 2001 [PubMed]
Escherichia coli Strain: O157:H7 (substrain RIMD 0509952) DDBJ     NCBI tax     NCBI entrez	49	5,498,450	Genome Atlas	5361	88%  1026 bp/gene	DNA Res. 8:11-22 February, 2001 [PubMed]

Link to a table with more E. coli Atlases.

Too Much Information!

Currently several hundred prokaryotic genomes have been sequenced, and nearly 100 genomes are publicly available for analysis. The flow of information is essentially the same as above, that is:

Genome -> Transcriptome -> Proteome

Link to a list of sequenced bacterial genomes

The DNA sequence contains several different types of information:

1. The DNA sequence can code for an amino acid sequence for proteins
• Directly - e.g., it is "easy" to predict protein sequence from DNA sequence.
• Indirectly - Scrambled genes in protozoa (changes at the DNA level)
• Indirectly - RNA editing
• Indirectly - RNA splicing
• Indirectly - Protein splicing (e.g., Inteins)



2. The DNA sequence can code for an RNA sequence
• tRNA
• rRNA
• snRNA
• telomeraseRNA
• other RNAs



3. The DNA sequence can code for protein binding sites



4. The DNA can code for architectural information
• intrinsic DNA curvature
• nucleosome positioning



5. The DNA can code for structural / stability information
• transcription initiation
• origins of replication
• mutational "hot spots"

REFERENCES

Photocopies of the following article will be provided for this lecture:

1. David W. Ussery,
   "Genome Databases",
   The Encyclopedia of Genetics, (Academic Press, New York, 2001), pages 517-521.        PDF file 
   Table with links to Genome Databases








Back to the CBS homepage

Back to Dave's Courses page

Last modified Wednesday, 4 September, 2002 by David Ussery