Note: Descriptions are shown in the official language in which they were submitted.
CA 02444604 2003-10-17
WO 01/54557 PCT/US01/12736
COMPUTATIONAL SUBTRACTION METHOD
Field of the Invention
The invention relates to a method and system for detecting microbes harbored
by a host
organism. In particular, the invention relates to a method and system for
detecting novel
infectious disease organisms associated with the pathogenesis of human
diseases.
Background of the Invention
Humans and animals are in continuous contact with microorganisms. Generally,
because
of the effectiveness of host defense mechanisms these microorganisms do not
cause disease.
However, some microorganisms (e.g., opportunistic pathogens) can become
infective in
particular types of individuals, such as those who are immunocompromised.
Still other
microorganisms are extremely virulent upon contact. For example,
microorganisms such as the
Ebola virus are associated with close to 100% fatality rates.
Traditional methods of correlating disease symptoms with the presence of a
microorganism rely on identifying through symptoms and/or through
epidemiological studies,
the likelihood that the disease is caused by an infectious agent and
attempting to culture
appropriate samples from the individual to isolate and identify the agent.
This can be
problematic where epidemiological evidence is unclear, particularly in the
case of pathogens
with long incubation periods (e.g., up to 10 years in the case of HIV and 20-
30 years in the case
of Mycobacterium leprae).
Even where epidemiological evidence suggests an infectious cause for a
disease, the
microorganisms responsible for these diseases can evade detection. For
example, Whipple's
disease, a debilitating disease associated with diarrhea and weight loss, was
for many years
described as "intestinal lipodystrophy" because no microorganism could be
cultured from
samples from patients with disease. However, the microbial origin of Whipple's
was suggested
by the dramatic response of patients to antibiotics and the presence of
bacilli observed in electron
micrographs of affected tissues. Still, the identification of the infectious
organism as an
actinomycetes awaited the advent of molecular techniques such as PCR. See,
e.g., Maiwald et
al., Clin. Infect. Dis. 32~: 457-463 (2001). PCR amplification of conserved
ribosomal
sequences also led to the detection of another unculturable bacteria, the
causative agent of
bacillary angiomatosis which is associated with the proliferation of small
blood vessels in the
WO 01/54557 CA 02444604 2003-io-17 pCT/USO1/12736
skin and visceral organs of patients with AIDS (see, Relman et al., New Engl.
J. Med. ~: 1573-
1580 (1990)). Another molecular technique, the DNA subtractive cloning method -
representational difference analysis (Lisitsyn, Trends Genetics Il: 303-307
(1995)), enabled the
discovery of the herpesvirus causing Kaposi's sarcoma (Chang et al., Science
265: 1865-1869
( 1994)).
A high throughput approach to identifying infectious organisms has been
described by
Cummings and Relman, Emerg. Infect. Dis. ~: 513-25 (2000). Cummings and Relman
report
using a DNA microarray comprising sequences from known pathogens to detect the
presence
pathogens in patient samples. However, the method will only be able to detect
pathogens for
which at least some sequence information is known.
Summary of the Invention
There is a need in the art to provide a systematic approach for the detection
and
identification of microbes which are harbored within a host organism,
particularly those
associated with pathogenesis. Therefore, in one aspect, the invention provides
a method of using
a computer system to identify a microbe inhabiting a host organism which
comprises the steps of
obtaining sequence information from a plurality of sequences from at least one
host organism
and searching a database of host organism genomic sequences to determine the
presence or
absence of the plurality of expressed sequences in the database. The absence
of at least one of
the sequences in the database indicates that the at least one sequence is a
candidate microbe
sequence. Individual sequences can be searched sequentially; however,
preferably, sets of
sequences are searched at a time.
In one aspect, the method comprises the steps of obtaining sequence
information from a
library of genomic DNA from a host organism and searching a database of
genomic sequences
from host organisms to determine the presence or absence of a sequence in the
library in the
database. A sequence that is present in the library but is absent in the
database is identified as a
candidate microbe sequence.
The microbe can be a symbiotic organism, such as a mutualistic organism, a
commensal
organism or a parasitic organism. The microbe can also be a pathogen. Microbes
which can be
identified by the method include, but are not limited to, phage, bacteria,
viruses, protozoa and
fungi. The host organism can be a microorganism, a plant, or an animal, such
as a mammal (e.g.,
a human being). The host organism can also be an insect, bird, or a fish.
2
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
In one aspect, the pluraliny of sequences from the least one host organism
comprises
expressed sequences. For exam ale, the plurality of sequences can comprise EST
and/or cDNA
sequences. Sequence informatic:n relating to expressed sequences can be
obtained by sequencing
a library of expressed sequences from one or more host organisms.
Additionally, or
alternatively, expressed sequence information can be obtained from a database
of expressed
sequences, such as an EST or cDNA database.
In one aspect, sequences from the at least one host organism suspected
harboring a
microbe are enriched for sequences which are present in the at least one host
organism and which
are not present in a plurality of host organisms which do not harbor the
microbe. Enrichment can
be performed using a subtractive hybridization assay, which can be a
differential gene expression
assay. Subtractive hybridization assays include, but are not limited to,
representational
difference analysis, SAGE, and suppression subtraction analysis.
Alternatively, enrichment can
be performed by electronically subtracting sequences from the at least one
host organism which
are stored in a first database from sequences of the plurality of organisms
which are stored in a
second database. In one aspect, the first and second databases are both
expressed sequence
databases and electronic subtraction is used to enrich for differentially
expressed sequences
which are expressed in the at least one host organism suspected of harboring a
microbe and not
expressed in the plurality of organisms which do not harbor the organism.
In one aspect, enriched sequences are then compared to sequences in a host
organism
genomic database to identify sequences in the at least one host organism
suspected of harboring
a microbe which are not present in the host organism genomic database. These
sequences are
identified as candidate sequences belonging to a microbe.
In a further aspect, one or more of the following sequences are eliminated
from the host
organism genomic database: vector sequences, mitochondrial sequences,
repetitive sequences,
sequences from other species, low quality sequences, known host organism mRNA
sequences,
and combinations thereof.
In a preferred aspect, the method according to the invention is used to
identify the
sequence of a pathogen. In this aspect, the at least one host organism is an
organism which has a
pathogenic condition, and sequences from the host organism (expressed or
genomic) are
compared to genomic sequences in a database from host organisms which do not
have the
pathogenic condition. The pathogenic condition can be a disease selected from
the group
consisting of an inflammatory disease, an autoimmune disease, and a cell
proliferative disease.
More particularly, the disease can be selected from the group consisting of:
sarcoidosis,
WO 01/54557 CA 02444604 2003-io-17 pCT/LJSO1/12736
inflammatory bowel disease (e.g., such as Crohn's disease), atherosclerosis,
multiple sclerosis,
rheumatoid arthritis, type I diabetes mellitus, lupus erythematosus, Hodgkin's
disease, and
bronchioalveolar carcinoma. Sequences from the at least one host organism
which do not match
sequences in the genomic database identified as candidate sequences belonging
to a pathogenic
organism. In one embodiment, the pathogenic organism is an infectious disease
organism.
In a further aspect, the invention provides a method of using a computer
system to
identify a microbe inhabiting a host organism, comprising the steps of:
obtaining sequence
information from a plurality of expressed sequences from at least one host
organism; and
searching a database of host organism genomic sequences to determine the
presence or absence
of the plurality of expressed sequences in the database, wherein the absence
of an expressed
sequence in the database identifies the expressed sequence as a candidate
microbe sequence.
Preferably, the plurality of sequences are from a library. Still more
preferably, the library is a
library of expressed sequences. In one aspect, the library comprises human
sequences. In
another aspect, the library comprises human sequences from one or more humans
having a
disease. The disease can be selected from the group consisting of an
inflammatory disease, an
autoimmune disease, and a cell proliferative disease. In one aspect, the
disease is selected from
the group consisting of sarcoidosis, inflammatory bowel disease,
atherosclerosis, multiple
sclerosis, rheumatoid arthritis, type I diabetes mellitus, lupus
erythematosus, Hodgkin's disease,
and bronchioalveolar carcinoma.
In still a further aspect, the invention provides a method of using a computer
system to
identify a microbe inhabiting a host organism comprising the steps of:
obtaining expressed
sequence information from a plurality of sequences from at least one non-
microbial host
organism and searching a database of microbial sequences to determine the
presence or absence
of the plurality of expressed sequences in the database, wherein the presence
of an expressed
sequence in the database identifies the expressed sequence as a candidate
microbe sequence. In
one aspect, the plurality of sequences are from a library of expressed
sequences. In another
aspect, the library of sequences comprises sequences from one or more humans
having a
pathological condition, e.g., such as an infectious disease.
Candidate sequences can be used as query sequences to search a database of
microbial
sequences, such as a database comprising bacterial and/or viral sequences.
Candidate sequences
also can be used to search databases comprising fungal sequences, parasitic
sequences, and/or
protozoan sequences. Candidate sequences also can be used as query sequences
to search a non-
redundant expressed sequence database comprising sequences from host
organisms.
4
W~ 01/54557 CA 02444604 2003-10-17 pC'j'/jJS01/12736
Candidate sequences or their complements can be used to probe a library of
sequences
from at least one microbe to identify first hybridizing sequences, preferably
sequences which are
longer in length (e.g., numbers of bases) than the candidate sequence.
Hybridizing sequences
can in turn be used to identify second hybridizing sequences which are longer
in length than the
first hybridizing sequences. Overlapping sequences which are identified can be
used to map the
genomic structure of the microbe. In some aspects, the complement of the
candidate sequence is
hybridized to RNA from the microbe and used to generate cDNAs.
The candidate sequence can be used to express a peptide; for example, by
operably
linking the candidate sequence to a promoter sequence in an expression vector.
Alternatively, or
additionally, sequences identified by probing a library of sequences using the
candidate sequence
as a probe can be used to express one or more peptides. Preferably, the
peptides are antigenic.
Still more preferably, the peptides can be administered to a host organism to
elicit a protective
immune response. Nucleic acid sequences expressing the peptides can also be
administered to
the host organism to elicit a protective immune response to the peptides
expressed by these
sequences.
The candidate sequence and/or other sequences identified by the candidate
sequence can
be used to detect the presence or absence of the microbe in a sample from the
host organism. For
example, the hybridization of the candidate sequence and/or the other
sequences to nucleic acid
sequences in the sample from the host organism under stringent conditions can
provide an
indication of the presence of the microbe in the sample. In preferred
embodiments, where the
microbe is a pathogen, detection of hybridization is used to provide a
diagnosis that the host
organism is infected by the pathogen.
Peptides expressed by the candidate sequences and/or sequences identified
using the
candidate sequence can be used as antigens to generate antibodies which can
also be used in
diagnostic assays. For example, in one embodiment, an antibody which
specifically binds to a
peptide expressed by the candidate sequence and/or sequences identified using
the candidate
sequence is contacted with a sample from the host organism and binding of the
antibody to a
polypeptide within the sample provides an indication that the host organism
harbors the microbe.
In some embodiments, the complementary sequence of a coding sequence of the
candidate sequence or of another sequence identified by the candidate sequence
is administered
to a host organism harboring the microbe in an amount sufficient to prevent
the expression of a
polypeptide encoded by the candidate sequence or the sequence identified by
the candidate
5
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
sequence in the host organism. The complementary sequence can further comprise
a cleaving
moiety for cleaving RNA (e.g., the complementary sequence can be a ribozyme).
In one aspect, a system for performing the method is provided. The system
comprises a
a first database comprising sequences from at least one host organism
suspected of harboring a
microbe and a second database comprising genomic sequences from host organisms
not
suspected of harboring the microbe. The system further comprises an
information management
system comprising a search and subtraction function for identifying sequences
in the first
database which are not present in the second database. In a preferred
embodiment, the
information management system comprises a sequence alignment function and can
compare a set
of sequences in the first database with all sequences in the second database.
The system
preferably comprises at least one user device connectable to the network and,
preferably, a high
speed, linear array processor.
In one aspect, the system comprises a program capable of implementing an
algorithm for
simultaneously comparing a plurality of sequences in a first database with all
sequences in a
second database, e.g., such as the algorithm implemented by the MEGABLAST
program.
However, in another aspect, the system comprises a program which sequentially
compares a
plurality of individual sequences from the first database with a plurality,
and preferably all,
sequences in the second database. Preferably, the system generates a result
sequence set
comprising sequences in the first database which do not match sequences in the
genomic
database.
In one aspect, the system comprises an identity or scoring matrix which
requires a score
of greater than or equal to 60 (e.g., equivalent to thirty identical
consecutive nucleotides). In
another aspect, the system iteratively computes the degree of alignment
between sequences in the
first and second database. Iterative computing preferably is performed using
progressively
smaller word sizes. In still a further aspect, the system provides one or more
programs for
performing one or more electronic subtraction functions for eliminating any
of: vector
sequences, repetitive sequences, mitochondrial sequences, sequences from non-
host organisms,
low quality sequences, known host organism mRNA sequences, and combinations
thereof, from
the genomic database.
The invention additionally provides a computer program product comprising a
computer
readable memory on which is embedded one or more programs for implementing any
of the
system functions and/or methods described above.
6
WO 01/54557 CA 02444604 2003-10-17 pCT/US01/12736
Brief Description of the Drawings
The objects and features of the invention can be better understood with
reference to the
following detailed description a~ id accompanying drawings.
Figure 1 is a flow chart demonstrating a method of computational subtraction
analysis
according to one embodiment of the invention to identify microbes harbored by
a human being.
Figure 2 is a schematic of a system according to one aspect of the invention
for
performing a computational subtraction analysis.
Detailed Description
The invention provides a method and system for performing computational
subtraction to
detect microbes harbored by a host organism. In some aspects, the microbes are
pathogens and
the system is used to identify sequences belonging to these pathogens which
can then be used in
methods of diagnosis and treatment. Alternatively, the microbes can be
symbiotic organisms,
such as commensal or parasitic organisms. Preferably, candidate sequences
identified as
belonging to a microbe are used to isolate and clone additional sequences from
the microbe.
Definitions
The following definitions are provided for specific terms which are used in
the following
written description.
As used herein, the term "expressed sequence" is a sequence which is
transcribed.
"Expressed sequence information" refers to the nucleotide sequence of an
expressed sequence
such as an RNA molecule, a cDNA molecule or a portion of genomic DNA which
corresponds
to an expressed sequence, e.g., such as those portions of a gene whose
complement will become
part of an RNA transcript. An expressed sequence may include both coding
sequences (i.e.,
codons which are translated into polypeptide sequences) as well as non-coding
sequence (i.e.,
untranslated sequences).
As used herein, "a match" between sequences refers to a level of sequence
similarity
equivalent to a BLAST score ranging from 40 (the equivalent of 20 consecutive
identical
nucleotides) to 2000 (the equivalent of 1000 consecutive identical
nucleotides)..
7
WO 01/54557 CA 02444604 2003-10-17 pCT/USO1/12736
As used herein, a query sequence is "present" in a database if the database
contains a
sequence which matches the query sequence and is "absent" in a database if the
database does
not contain the matching sequence.
As used herein, a "low quality sequence" is a sequence which has greater than
2.5% N
nucleotides, i.e., nucleotides whose identity cannot be determined at 95%
confidence levels.
As used herein, "symbiosis" or a "symbiotic relationship" refers to an
association
between two organisms that live together. Symbiotic relationships include
mutualistic
relationships, commensalistic relationships, and parasitic relationships.
As used herein, "mutualism" or a "mutualistic relationship" refers to a
mutually-
beneficial association between two organisms.
As used herein, "commensalism" or a "commensalistic relationship" refers to an
association between two organisms where one organism may benefit but neither
is harmed.
As used herein, "parasitism" or a "parasitic relationship" refers to an
association between
two organisms in which one organism lives at the expense of the other organism
and can cause
damage to the other organism.
As used herein, a "pathogen" is an organism that can cause disease in another
organism
(e.g., the host organism).
As used herein, a "microbe" is any organism that can live and/or replicate
within a host
organism for at least a portion of its life cycle. While some microbes can
exist for at least a
portion of their life cycle intracellularly within the cells of a host
organism, microbes which
grow and/or replicate extracellularly are also encompassed within the scope of
the invention.
Microbes include, but are not limited to, phage, viruses, gram-positive and
gram-negative
bacteria, protozoa, small unicellular and multicellular eukaryotes (e.g.,
fungi, such as yeast), and
the like. The term "microbe" and "microorganism" are used interchangeably
herein.
As used herein a "host organism" can any organism that can harbor (e.g.,
provide a
habitat and/or nutrients for) another organism. Thus, a host can be a bacteria
which harbors a
phage, a simple eukaryote such as yeast which can harbor a bacteria, or a
mammal such as a
human being which can harbor by any of the foregoing.
As used herein, "infection" refers to the growth of a pathogen in a host
organism.
8
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
As used herein, an "infectious disease" refers to a disease that can be
transmitted from
host organism to host organism.
As used herein, a "carrier" refers to a patient who shows full recovery after
infection and
displaying symptoms but still carries and is capable of spreading the
infectious form of a
microbe.
As used herein, "a sequence identified by a candidate sequence" refers to
genomic
sequences of microbes to which the candidate sequence or its complement
hybridizes, or to
which the latter genomic sequences hybridize, under stringent conditions. In
some
embodiments, sequences are identified by the candidate sequence
electronically, e.g., by
searching a database of sequences from one or more microbes. Sequences which
are identified
as belonging to the same microbe as the organism from which the candidate
sequence was
obtained are said to be "identified by the candidate sequence."
As used herein, "stringent conditions" refer to conditions under which a
sequence will
specifically bind to its complement to enable detection of the complement and
to distinguish the
I S complement from other nucleic acid sequences in a sample. Stringency
conditions are described
in Sambrook et al., In Molecular Cloning: A Laboratory Manual, 2"'~ edition,
vols. 1-2. Cold
Spring Harbor Press ( 1989), the entirety of which is incorporated by
reference herein. As used
herein, stringent conditions require at least 80% base pairing, more
preferably, at least 90-95%
base pairing, and most preferably, at least 98% base pairing.
As used herein, a "fragment" of a candidate sequence or a sequence identified
by the
candidate sequence refers to a sequence which is shorter in length than the
candidate sequence
but sufficiently long to specifically hybridize to the candidate sequence. In
one embodiment, a
fragment ranges in size from 6 nucleotides to one less nucleotide than the
full-length sequence.
As defined herein, the "a promoter operably linked" to another sequence refers
to a
promoter and/or promoter element andlor enhancer elements) capable of
inducibly or
constitutively causing transcription of the other sequence.
As used herein, a "bodily fluid" refers to any of blood, plasma, sera, urine,
CSF fluid,
sputum, breast exudates, pus, and the like.
As used herein, "computational subtraction" or "electronic subtraction" or
"filtering"
refers to a computational method of eliminating records (e.g., such as
sequences) from a
database.
9
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
Computational Subtraction
In one aspect, the invention provides a systematic method to identify
sequences of
microbes capable of inhabiting a host organism. The microbes can be pathogenic
and associated
with an infectious disease. However, the microbes can also exist symbiotically
within a host
organism, e.g., in a mutualistic, commensal, or parasitic relationship within
the host organism.
The microbe can be any of a phage, a virus (e.g., an RNA or DNA virus), a
bacteria, a protozoa,
or other microorganism, a small unicellular or multicellular eukaryotic
organism (e.g., a fungus,
such as yeast), and the like. The host organism can be a microorganism, a
fungus, an animal, or
a plant. Preferably, the host organism is a mammal, such as a human being or a
domestic animal.
However, the host organism can also be an insect, bird, or a fish. Host
organism sequences can
be obtained from particular tissues or cells of the host organism, or from
cell lines derived from
these tissues or cells, or from bodily fluids from the host organism.
The invention provides a computational subtraction method for detecting and
identifying
microbe sequences. The method comprises comparing the sequence information of
a plurality of
sequences obtained from one or more host organisms with sequences in a genomic
database of
host sequences to identify which of the plurality of sequences are not found
(i.e., do not match
other sequences) in the database. Sequences which are not found in the
database are identified as
candidate sequences which are likely to belong to a microbe. Preferably,
sequence information
from sets of sequences (two or more sequences, and preferably ten or more
sequences) are
compared against the entire genomic database at a time.
Any number of nucleotide sequence alignment algorithms can be used for this
purpose,
including those known in the art. For example, in one aspect, the algorithm of
Zhang et al., J.
Comput. Biol. 7 1-2 : 203-14 (2000), which is embodied in one form in the
MEGABLAST
program, is used to compare sequences in an entire database of sequences from
one or more host
organisms (a "test database") against a genomic database. Smaller sets of
sequences (e.g., at
least two or at least ten) can also be compared. In some aspects, sequences
from the plurality of
sequences can be compared sequentially, individually, against genomic
databases, e.g., such as
by using the BLAST program described in Altschul et al., J. Mol. Biol. 21 S:
403-410 ( 1990);
Madden et al., Meth. Enzymol. 266: 131-141 (1996); and Zhang et al., Genonze
Res. 7: 649-656
(1997), the entireties of which are incorporated by reference herein. Other
programs whose goal
is sequence similarity searching also can be used, such as FASTA, SSAHA, or
any type of word
hashing program such as is known in the art (see, e.g., Pearson, Proc. Natl.
Acad. Sci. USA
W~ 01/54557 CA 02444604 2003-10-17 pCT/US01/12736
8~: 2444-2448 (1988); Leung rt al., J. Mol. Biol. 221 4 : 1367-1378 (1991),
the entireties of
which are incorporated by referer ce herein).
Methods of determining tae significance of sequence alignments are known in
the art and
are described in Needleman and '.?Vunsch, J. ofMol. Biol. 48: 444 (1970);
Waterman et al., J.
Mol. Biol. 147: 195-197 (1980); Karlin et al., Proc. Natl. Acad. Sci. USA 87:
2264-2268 (1990);
Karlin et al., Proc. Natl. Acad. Sci. USA 90: 5873-5877 (1993); Dembo et al.,
Ann. Prob. 22:
2022-2039 ( 1994) and Altschul, In Theoretical and Computational Methods in
Genome
Research. (S. Suhai, ed.), pp. 1-14, Plenum, New York; the entireties of which
are incorporated
by reference herein.
In some aspects, the genomic database is searched for short perfect matches of
a set
length (i.e., a word size). This enables a more rapid comparison than
window/stringency
matching. In one embodiment, a word size ranging from 10-30 bases is used.
Preferably, a
series of sequential searches is performed, using progressively smaller word
sizes ranging from
30 to 10 bases. More preferably, a first search using a word search of 24 is
performed, followed
by a second word search of 20, followed by a third word search of 16, followed
by a fourth word
search of 12. In one aspect, a test sequence is shifted to the left or right
of sequences in the
database to identify maximal regions of alignment.
In some aspects, a scoring matrix is used to identify the likelihood that one
or more
sequences in the test database do not match or are absent from the genomic
database. Preferably,
scores of greater than or equal to 60 are required. In one aspect, the scoring
matrix assigns a
match if there is a BLAST score ranging from 40 (the equivalent of 20
consecutive nucleotides)
through 2000 (the equivalent of 1000 consecutive nucleotides). In another
aspect, a matrix is
used which assigns expectation values to matches and mismatches after
alignment. Expectation
values can be adjusted to require that a score does not grow simply by
extending the alignment in
a random way. For example, in one embodiment, expectation values of from 10-
Z° - 10-3 can be
selected, and preferably, expectation values of 10-7 are used. Gap values can
be set to any
desired value as is routine in the art (see, e.g., Smith et al., 1981, J. Mol.
Evol. 18(1): 38-46,
Levitt et al., 1998, Proc. Natl. Acad. Sci. USA 95(11): 5913-5920, the
entireties of which are
incorporated herein by reference.
In one aspect, a results database is created, preferably comprising sequences
from the test
database which are ranked according to their alignment with sequences in the
genomic database.
Preferably, sequences which show a high degree of alignment to genomic
sequences from the
11
WO 01/54557 CA 02444604 2003-10-17 pCT/US01/12736
host organism (e.g., having at least 20-1000 consecutive identical
nucleotides) are not included
in the results database or are subsequently removed from the results database.
In a preferred embodiment, as shown in Figure 1, a subtraction operation is
performed to
remove sequences from either the genomic database and/or the test database
and/or the results
database. For example, subtraction operations can be used to remove vector
sequences,
repetitive sequences, mitochondria) sequences, sequences from other species,
low quality
sequences, known host organism mRNA sequences, and the like. It should be
obvious to those
of skill in the art that the order of subtraction operations is not critical
and that one or more
subtraction operations can be used. In certain aspects, after filtering
operations to filter sets of
candidate sequences through one or more of a vector sequence database, a
repetitive sequence
database, a mitochondria) sequence database, a non-host species database,
and/or a known host
organism mRNA database ("filtering databases"), a first candidate sequence set
of sequences is
again compared to the host organism genomic database, and/or one or more
filtering databases
using a reduced word size than was used in the previous series of operations,
to generate a
second candidate sequence set which is then stored in a results database. In a
preferred aspect,
low quality sequences are removed, before or after filtering.
Test Database
In one aspect, the test database is an expressed sequence database of
sequences from the
host organism, such as an EST or cDNA database (e.g., a library database).
Such databases are
known in the art and include, but are not limited to, human expressed sequence
databases such as
the NCBI EST database, the LIFESEQTM, database (Incyte Pharmaceuticals, Palo
Alto, Cali~),
the random cDNA sequence database from Human Genome Sciences, the EMEST8
database
(EMBL, Heidelberg, Germany), and the like (see, also, Boguski et al., 1993,
Nat. Genet. 4(4):
332-333, the entirety of which is incorporated by reference herein).
The test database also can be generated by inputting and storing sequence
information
obtained by sequencing a plurality of nucleic acids from a library of
expressed sequences from
one or more host organisms suspected of harboring a microbe, into a user
device of a system 1
(shown in Figure 2) as described further below. Libraries of expressed
sequences can be
generated using total RNA or polyadenylated RNA, and by using random priming
or oligodT
priming or a combination of these methods. Such techniques are known in the
art. Libraries of
particular interest include, but are not limited to, libraries of expressed
sequences from one or
more patients with an inflammatory disease, an autoimmune disease, and a cell
proliferative
disease. For example, in one aspect, libraries of expressed sequences from one
or more patients
12
WO 01/54557 CA 02444604 2003-io-17 pCT/USO1/12736
with a disease selected from the group consisting of sarcoidosis, inflammatory
bowel disease
(such as Crohn's disease), atherosclerosis, multiple sclerosis, rheumatoid
arthritis, type I diabetes
mellitus, lupus erythematosus, Hodgkin's disease, and bronchioalveolar
carcinoma are used to
obtain expressed sequence information. Preferably, the creation of such
libraries is performed to
minimize manipulation of tissue (e.g., by careful attention to sterility and
avoidance of
amplification methods) to avoid spurious contamination of such libraries with
bacterial
sequences.
While the test database can consist of entirely expressed sequences, the test
database can
also be a genomic sequence database. For example, the test database can
comprise sequence
information from a plurality of sequences in a genomic library from one or
more host organisms
suspected of harboring a microbe. Preferably, genomic sequence test databases
are used to
identify expressed sequences of microbes which are not polyadenylated (and/or
which have
integrated into the genome of the host organism), e.g., such as bacterial
expressed sequences
which would likely escape detection in expressed sequence libraries generated
from
polyadenylated RNA.
The test database can be enriched for sequences which are found in host
organisms)
suspected of harboring a microbe and which are not found in host organisms not
harboring the
microbe. In one aspect, the enrichment method comprises combining genomic test
sequences
with other genomic sequences (reference sequences), expressed test sequences
with expressed
reference sequences, or expressed test sequences with genomic reference
sequences, and
removing sequences which are common to both test and reference sequence sets,
thereby
enriching for test sequences which are not found in a reference set of
sequences.
For example, in one aspect, a subtractive hybridization method is used to
enrich for
expressed sequences in a sample of nucleic acids from a host organism which is
suspected of
harboring a microbe and which are not expressed in host organisms which do not
harbor the
microbe. Samples can comprise total nucleic acids, polyadenylated RNA, or
total RNA.
Subtractive hybridization methods to enrich for differentially expressed
sequences are known in
the art and include, but are not limited to, SAGE (Serial Amplification of
Gene Expression) (see,
e.g., Velculescu et al., Science 270: 484 (1995) and U.S. Patent No.
5,866,330), subtractive
hybridization of cDNA libraries (e.g., using magnetic beads, as described in
WO 97/07244 Al),
cDNA representational difference analysis (e.g., Hubank and Schatz, Nucl.
Acids Res. 22: 5640-
5648 (1994)), and suppression subtraction analysis (see, e.g., U.S. Patent
5,565,340).
Subtractive hybridization methods can also be used to enrich for sequences
which are present at
13
WO 01/54557 CA 02444604 2003-io-17 pCT~JS01/12736
different levels in different populations of genomic DNA. Such methods
include, but are not
limited to, representational difference analysis, such as described in U.S.
Patent No. 5,958,738
and CLONTECH's PCR-SelectTM Bacterial Genome Subtraction technique (see, e.g.,
Diatchenko et al., Proc. Natl. Acad. Sci. USA 93: 6025-6030 (1996);
CLONTECHniques ~:
2-5 (1995)). The entireties of these references are incorporated by reference
herein.
In another aspect, enrichment is performed electronically. For example,
sequences from
at least one host organism suspected of harboring a microbe stored in a test
database can be
subtracted from sequences in a "reference database" comprising sequences from
a plurality of
host organisms not harboring the microbe. In one aspect, the test database and
reference
database are both expressed sequence databases and electronic subtraction is
used to enrich for
differentially expressed sequences which are expressed in the at least one
host organism and
which are not expressed in the plurality of host organisms. Methods for
electronic subtraction
analysis of expressed sequences are described in U.S. Patent 6,114,114, for
example, the entirety
of which is incorporated by reference herein.
In some embodiments, the test database is a relational database which
segregates
particular types of sequences from other types of sequences within the
database. For example, in
one aspect, expressed sequence information can be subdivided within an
expressed sequence
database according to a particular tissue, or cell type, or cell line, in
which the sequence is
expressed. In these embodiments, particular portions of the test database can
be compared to the
genomic database during a search, sequentially, or simultaneously.
In one aspect, once a candidate sequence is identified, it is compared to a
nucleotide
sequence database comprising sequences from a plurality of species, to
identify the microbial
organism genus to which the sequence belongs or to which the species is
related evolutionarily.
For example, GenBank's nucleotide or "nt" database can be used. In another
aspect, the
candidate sequence can be used as a query sequence to search a database
comprising only
microbe sequences. In one embodiment, the database is a microbial sequence
database which
can be a viral sequence database, or a fungal or parasite sequence database.
Such databases are
known in the art and include, but are not limited to, the Incyte Microbial
Database, the TIGR
Microbial Database, the TIGR Parasites Database, TIGR Fungal Database, and the
TIGR Viral
Genome Sequencing Project Database. This step can be used to identify or
evaluate the
taxonomic relationship between the candidate sequence and sequences of other
known microbes
for which genomic sequence information is known.
14
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
In still another aspect, candidate sequences are compared to sequences in a
non-
redundant RNA database to determine whether the sequence matches known host
organism RNA
molecules. In still a further asl;ect, a candidate sequence is conceptually
translated to identify
open reading frames and the ammo acid sequences of a polypeptide encoded by
the candidate
sequence can be used as a query sequence to search a protein sequence database
comprising
microbial sequences (i.e., the database can comprise multiple species
sequences in addition to
microbial sequences, such as the GenBank nr database, or the database can
comprise exclusively
microbial sequences). Preferably, the sequence is also used as a query
sequence to search a
nucleotide sequence database comprising microbial sequences (e.g., such as the
GenBank nt
database or an exclusively microbial sequence database) to identify sequence
whose conceptual
translations match known microbial proteins but whose nucleotide sequences do
not match
microbial nucleotides. These latter classes of sequences, which are preferably
stored in a results
database, are likely to identify sequences belonging to microbes of the same
genus as the
microbe whose protein was identified as a match, but which do not necessarily
represent
microbes belonging to the same species, i.e., the sequences are likely to
represent previously
uncharacterized microbes.
Genomic Databases
Genomic databases for a variety of host organisms are also known in the art,
and include,
but are not limited to, the NCBI GenBank database (see, e.g., http://
www.ncbi.nlm.nih.gov/
entrez/query.fcgi?db=Genome), the Celera Human Genome (http://www.celera.com),
the
Genetic Information Research Institute (GIRD (http://www.girinst.org) and
Human Genome
Fragment database, TIGR databases (e.g., the TIGR Human Gene Index Database),
and the like.
The genomic database also can be generated by inputting and storing sequence
information obtained by sequencing a plurality of nucleic acids from a genomic
library of
sequences from one or more host organisms which do not harbor microbes, into a
user device of
a system described further below.
Genomic databases contemplated according to the invention include genomic
sequence
information from any of the host organisms described above, e.g., from a
microorganism, a
fungus (e.g., yeast), an animal (insect, bird, fish, or mammal, such as a
human being or domestic
animal) or a plant.
WO 01/54557 CA 02444604 2003-io-17 pCT/USO1/12736
Svstem For Performing The Computational Subtraction Method
The invention further provides a system 1 for performing the computational
subtraction
analysis described above (see, Figure 2). In one aspect, the system 1
comprises a first database 2
(e.g., the test database) comprising sequences from at least one host organism
suspected of
harboring a microbe and a second database 3 comprising genomic sequences from
host
organisms not harboring the microbe. The system 1 further comprises an
information
management system 4 comprising a search function for identifying sequences in
the first
database 2 which are not present in the second database 3. In a preferred
embodiment, the
system 1 further comprises a program embodied in a computer readable medium
for executing
sequence alignments between at least a first sequence in the first database
and a plurality, and
preferably, all sequences in the second database. The program can be part of a
server 5 (which
also can store program applications required by the information management
system 4) or part of
a processor which is part of a user device 9. Preferably, however, the user
device 9 is in
communication with the server 5 and/or other servers (not shown). The user
device 9 can be a
computer, a laptop, a wireless device, and the like. The system further can
include additional
user devices 9, output devices 6 (e.g., printers), and input devices (e.g.,
keyboards 7, mice,
joysticks, and the like). The user device 9 preferably includes an interface 8
which can be
displayed by the device 3 in response to the user accessing the system 1 to
activate the
information management system 5. The system 1 is preferably connectable to the
network 10,
enabling a user to access the system remotely from any user device 3 that is
connectable to the
network 10.
In preferred embodiments, sets of sequences (at least 2, at least 10, at least
100, or at least
500) in the test database 2 are compared at a single time with sequences in
the genomic database
3. In one embodiment, the information management system 5 comprises a program
which is
capable of implementing an algorithm, such as the one used in the MEGABLAST
program for
performing this function (see, e.g., Zhang et al., supra). However, in other
embodiments, the at
least 2, at least 10, at least 100, or at least 500 sequences are compared
individually and
sequentially with sequences in the genomic database 3.
In preferred embodiments, the user device 3 or the server 5 (or another host
computer)
comprises a high speed, linear array processor that can locate highly similar
sequence segments
(e.g., having a BLAST score of at least 40) from any at least two sequences.
In one aspect, the
processor comprises a high speed circuit chip that provides an equivalent of
about 400,000
16
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
transistors or 100,000 gates, as described in U.S. Patent No. 5,964,860, for
use in performing
high speed sequence analyses.
In one aspect, the system 1 further comprises an input device 7 that receives
a set of
sequences (either sequentially or simultaneously), a memory that stores the
set of sequences (not
shown), and a processor that transfers information from the set of sequences
to the memory (e.g.,
in the form of data characters representing nucleotide bases in the set of
sequences). The
processor can be part of a user device 3, but is preferably part of the server
S. In another aspect,
the system 1 further includes an identity matrix and a result sequence set
(e.g., from the results
database described above) (not shown), in which members of a set of compared
sequences are
ranked according to their degree of match to sequences in the genomic database
3. In a further
aspect, the results sequence set can include sequences which do not match
sequences in the
genomic database. Sequences which have the least amount of match (as
determined using
parameters established by the user) can be displayed on an interface 8 of the
user device 8 in
response to a user query to match sequences.
In one aspect, the identity matrix is pre-selected by the user to require a
match score of
greater than or equal to 60 with a word size of between 10 and 30. In one
embodiment, the
system 1 iteratively computes the degree of alignment between sequences using
progressively
smaller word sizes from 30 to 10, (e.g., first using a word size of 24, then a
word size of 20, then
a word size of 16, then a word size of 12). Preferably, the score value
remains the same and is
some value greater than or equal to 60. The matrix is designed to eliminate
low quality
sequences (e.g., as determined using a base calling program such as PHRED),
short sequences
(less than 150 nucleotides), or sequences comprising a maximum number of
ambiguous or
unreadable nucleotides, such that there is a minimum length of quality
sequences (e.g.,
sequences whose bases have a high confidence (at least 95%) of being accurate)
of at least 50
nucleotides, and preferably at least 150 nucleotides.
In one aspect, the system 1 provide one or more programs for performing one or
more
electronic subtraction functions analogous to an electronic subtractive
hybridization. For
example, in one aspect, the system 1 is capable of eliminating, in response to
a user command or
in response to a pre-programmed set of instructions, any o~ vector sequences,
repetitive
sequences, mitochondria) sequences, sequences from other species, low quality
sequences,
known host mRNA sequences (i.e., sequences known to belong to the host
organism), and
combinations thereof.
17
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
In a further aspect, the invention provides a computer program product
comprising a
computer readable memory on which is embedded one or more programs for
implementing any
of the system 1 functions described above.
Methods of Using Candidate Sequences
Cloning and Sequencing Genomic DNA From Microbes
In one aspect, candidate sequences identified using the methods and system 1
described
above are used as probes to probe a library of sequences from at least one
microbe. The microbe
can be a phage, a virus, a bacteria, a protozoa, or other microorganism, a
small unicellular or
multicellular eukaryotic organism, such as a fungi (e.g., yeast), and the
like. Microbes can be
cultured from host organisms to provide nucleic acids suitable for generating
libraries using
methods known in the art. Preferably, the library is a genomic library.
Alternatively, where
microbes cannot be cultured, libraries can be generated from genomic or
expressed sequences
from which host organism sequences have been subtracted as described above. In
some aspects,
microbe sequences which are enriched in these samples can be ligated to
linkers or adapters and
amplified using primers which hybridize to these linkers or adapters.
Alternatively, or
additionally, the linkers or adapters can include promoter sequences and
microbe sequences can
be amplified by providing polymerases which recognize these sequences and the
appropriate
nucleotides (e.g., using a transcription-based amplification system). These
methods can be
complemented by additional rounds of computational subtraction as described
above by
sequencing enriched sequences and subtracting sequence information
corresponding to these
enriched sequences from a genomic database to identify enriched sequences
which are not found
in the genomic database.
In one aspect, candidate sequences are used to identify hybridizing sequences
within the
library which are longer in length than the candidate sequence, either at the
5' end or 3' end or
both. These longer sequences are used, in turn, to identify other sequences
which are preferably
longer in length either at the 5' end or 3' end or both. Overlapping clones
can be mapped using
restriction enzyme analysis in combination with Southern analysis, and/or
sequence analysis, to
further characterize the genome structure of the microbe. Preferably, genomic
sequence
information is inputted into a microbial genomic database (i.e., a database
comprising only
microbial sequences).
Microbe sequences can be evaluated using a sequence analysis program such as
the Gene
Locator and Interpolated Markov Modeler, or GlimmerTM , program to identify
coding sequences
18
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
and to distinguish such sequences from non-coding DNA (see, e.g., Salzberg et
al. Nucl. Acids
Res. 2:544-8 ( 1998). A versic in of Glimmer designed for small eukaryotes is
described in
Salzberg et al., Genomics 59: 24-: 1 (1999). The entirety of these references
is incorporated by
reference herein.
In one embodiment, RNA samples are obtained from host organisms harboring the
microbe (e.g., total RNA or polyA RNA if the microbe's RNA is polyadenylated)
and a
complement of the candidate sequence is used as a primer to generate cDNA
molecules from the
RNAs obtained. In one embodiment, cDNAs are generated using a RACE method
(see, e.g.,
Siebert et al., In Gene Cloning and Analysis by RT PCR (BioTechniques Books,
Natick, MA),
pp. 305-320 ( 1998); Don et al., Nucl. Acids Res. 19: 4008 ( 1991 ); Roux, PCR
Methods Appl. 4:
5185-5194 (1995); the entireties of which are incorporated by reference
herein) to identify the 5'
and or 3' end of a particular RNA transcript. Preferably, the sequences of
cDNA clones are
inputted into a results database for comparison to a database comprising
microbial nucleotide
sequences.
In some aspects, when a host organism has been identified, candidate sequences
and/or
their complements can be used as primers in PCR or RT-PCR assays to identify
additional
microbial sequences of interest, for example, in nucleic acids obtained from
cultures of these
microbes. In one aspect, asymmetric or one-directional PCR can be performed
using the
candidate sequence or its complement as a single primer in primer extension
reactions to identify
microbial sequences flanking the primer sequence in the microbial genome or in
a microbial
transcript. One-directional PCR is known in the art and is described in U.S.
Patent No.
6,184,025, for example, the entirety of which is incorporated by reference. In
other aspects, at
least two primers corresponding to the candidate sequence are used (e.g.,
primers capable of
amplifying a nucleic acid fragment which comprises a subsequence of the
candidate sequence of
at least 50 nucleotides). In addition to detecting microbial sequences
amplifiable using these
primers, the primers can be used to verify that the candidate sequences do not
represent
previously unsequenced host genomic DNA. For example, the primers can be used
in
amplification reactions with host genomic DNA to verify that no amplification
of host genomic
sequences occurs.
Diagnostic Methods
Candidate sequences, their complements, or sequences identified by candidate
sequences
(e.g., such as by any of the assays described in the preceding section), can
be used in
hybridization assays to detect the presence of a microbe in a sample. Although
such methods are
19
WO 01/54557 CA 02444604 2003-10-17 pC't/USO1/12736
described as "diagnostic", this does not imply that the method is necessarily
used to determine
the presence or absence of a pathogenic condition in an organism. For example,
diagnostic
methods can be used to detect the presence of a commensal microbe within a
sample, which can,
in some instances, be desirable (e.g., such as when the microbe produces
vitamins for the host).
In some instances, however, the hybridization assays can be used to detect the
presence of one or
more pathogens in a sample from an organism, and the results of such as assay
can be used to
provide treatment options for the organism. In still other aspects, the
hybridization assays are
used to detect carrier organisms which are infected by pathogens but which do
not show
symptoms of a pathogenic condition.
In one aspect, nucleic acids from a sample obtained from a host organism
(e.g., a cell, a
tissue sample, a bodily fluid, a lavage specimen, and the like) are contacted
under stringent
conditions with a test sequence derived from the candidate sequence. As used
herein, "a test
sequence derived from a candidate sequence" refers to the candidate sequence
itself, or a
fragment thereof, or another sequence from the microbe which the candidate
sequence has been
used to identify, or to complements of any of these sequences.
The test sequence can be used as a diagnostic probe to detect expressed
sequences or
genomic sequences of the microbes in the sample by detecting the formation of
a hybridization
complex between the test sequence and a nucleic acid in the sample. In one
embodiment, test
sequences are labeled with detectable labels. However, in other embodiments,
the test sequence
is bound to a molecule which is detectably labeled or which itself can bind to
detectably labeled
molecule(s). In one aspect, the amount of test sequences bound is used to
provide an indication
of the number of microbes in a sample (for example, by providing a comparison
to test samples
comprising a known amount of microbes). In another aspect, either the sample
sequences, or
probe sequences, or both, are amplified (e.g., by PCR, LCR or some other means
of
amplification) to increase the sensitivity of the assay. In still a further
embodiment, the test
sequence itself is used as a primer in an amplification assay or a reverse
transcription-based
assay. Methods of labeling, hybridizing, amplifying and quantitating nucleic
acids are known in
the art. Probes can be obtained by restriction digestion of cloned sequences
or can be
synthesized using means known in the art. PNA probes can also be used to
enhance the
specificity of assays.
In some embodiments, panels of nucleic acid sequences representing different
regions of
the genome of the microbe can be used simultaneously or sequentially to detect
the microbe. In
still another embodiment, panels of nucleic acid probes from different
microbes can be used in
WO 01/54557 CA 02444604 2003-10-17 pCT/USO1/12736
the diagnostic assays described above. The probes, or oligonucleotides
comprising probe
sequences, can be immobilized on a substrate (e.g., a microarray) as described
in Cummings et
al., supra, to increase the throughput of diagnostic assays.
In one aspect, the candidate sequence, or a sequence identified by the
candidate sequence,
is used to express a peptide, for example, by operably linking the candidate
sequence to a
promoter sequence in an expression vector. In some embodiments, the candidate
sequence is
linked in frame to a cleavable amino acid sequence whose expression is
operably linked to the
promoter sequence. Alternatively, a peptide can be synthesized using the
predicted amino acid
sequence of the candidate sequence or a coding sequence of the sequence
identified by the
candidate sequence. Preferably, the peptide is an antigenic peptide. The
peptide can be used to
generate antibodies which specifically bind to the peptide and to polypeptides
or proteins
comprising the peptide.
Methods of generating antibodies are know in the art and are described in
Kohler and
Milstein, Nature 256: 495-497 (1975); Kosbor et al., Immunology Today g: 72
(1983); Cote et
al., Proc. Natl. Acad. Sci. U.S.A. 80: 2026-2030 (1983), (Morrison et al.,
Proc. Natl. Acad. Sci.
U.S.A. 81: 6851-6855 (1985); Neuberger et al., Nature 312: 604-608 (1984);
Takeda et al.,
Nature 314: 452-454 (1985); and U.S. Patent No. 4,946,778, the entireties of
which are
incorporated by reference herein. Antibodies encompassed within the scope of
the invention
include, but are not limited to, monoclonal antibodies, polyclonal antibodies,
double chain
antibodies, single chain antibodies, chimeric antibodies, antibody fragments
comprising at least
one antigen binding site, and the like.
In one aspect, antibodies specific for peptides expressed by nucleic acids
from the
microbe are used in histological assays, such as immunohistochemistry,
immunofluorescence,
immunoelectron microscopy, and the like. However, antibodies can also be used
in
immunoassays as are routine in the art. The detection of binding of an
antibody to a sample from
a host organism suspected of harboring a microbe can be used to provide a
diagnosis that the
organism harbors the microbe (e.g., that the microbe may be found on or within
its cells, or in
bodily fluids from the organism). For example, the antibodies according to the
invention can be
used to detect microbes which are shed by host cells and which may be present
in bodily fluids
outside of cells or in proximity to cells or tissues from the host organism,
or to detect antigens
which are presented after processing of polypeptides of a microbe by host
cells (e.g., by host cell
MHC class I molecules), or to detect microbes which typically exist
extracellularly within a host
organism, such as bacteria.
21
W~ 01/54557 cA 02444604 2003-io-i7 pCT/US01/12736
As with the nucleic acid probes described above, panels of antibodies specific
for a single
microbe can be used as probes, either simultaneously or sequentially. Panels
of antibodies
specific for a plurality of microbes can also be used. In one embodiment,
antibodies are arrayed
on a substrate to increase the throughput of the analysis.
In some aspects, peptides themselves can be used as diagnostic reagents. For
example,
peptides can be reacted from sera from an organism suspected of containing a
microbe to detect
the presence of circulating antibodies which react with the peptides.
Antisense Nucleic Acid Molecules
In one aspect, the invention provides a sequence which is a complement or an
antisense
sequence of a coding sequence of the candidate sequence or of the coding
sequence of another
sequence which has been identified by the candidate sequence. The antisense
sequence can be
administered to a host organism in an amount sufficient to prevent the
expression of a
polypeptide encoded by the candidate sequence or the other sequence identified
by the candidate
sequence. Techniques of generating antisense constructs are known in the art
and are described
in, for example, Stein et al., Cancer Research 48: 2659-2668 ( 1988); Walden
Genes c~
Development 2: 502-504 (1988); Marcus-Sekura, Anal. Biochemistry 172: 289-295
(1988); Zon,
J. of Protein Chemistry 6: 131-145 (1987); Zon, Pharmaceutical Research 5: 539-
549 (1988);
and Loose-Mitchell, TIPS 9: 45-47 (1988); the entireties of which are
incorporated by reference.
Antisense nucleic acids according to the invention additionally can be
modified to enhance their
stability in vivo, as described in Agrarwal et al., Proc. Natl. Acad. Sci. USA
85: 7079 ( 1988), and
Sarin et al., Proc. Natl. Acad. Sci. USA 85: 7448 (1988), for example, the
entireties of which are
incorporated herein by reference.
Antisense nucleic acids also can be modified to include a cleaving agent for
cleaving a
molecule to which the antisense nucleic acid binds. For example, the nucleic
acid can be
engineered to sequences which provide the function of a ribozyme. Sequences
for use in
constructing ribozyme vectors are described in, for example, Rossi et al.,
Aids Research and
Human Retroviruses 8: 183 ( 1992); Hampel and Tritz, Biochernistry 28: 4929 (
1989); Hampel et
al., Nucleic Acids Research 18: 299 (1990); Perrotta et al., Biochemistry 31:
16 (1992); Guerrier-
Takada et al., Cell 35: 849, (1983); U.S. Pat. No. 4,987,071; Scanlon et al.,
PNAS 88: 10591-5
(1991); Dropulic et al., J Virol. 66: 1432-41 (1992); Weerasinghe et al., J
Virol. 65: 5531-5534
(1991); Ojwang et al., PNAS 89: 10802-10806 (1992); Chen et al., Nucleic Acids
Res. 20: 4581-
1589 (1992); and Sarver et al., Science 247: 1222-1225 (1992); the entireties
of which are
incorporated herein by reference.
22
WO 01/54557 CA 02444604 2003-io-17 pCT/US01/12736
Antisense molecules can be administered directly to a target site. For
example, antisense
molecules can be administered topically (e.g.., to skin), by direct injection
into cells (e.g., such as
tumor cells), by direct administra..ion to a tissue which has been exposed by
surgery, or through a
medical access device, such as a ~:athetc~r or endoscope, which can deliver
the molecule directly
to the target site (e.g., by bringing the tissue into contact with a solution
comprising the antisense
molecules). In another aspect, antisense molecules are administered to the
patient enterally or
parenterally. Antisense molecules can be administered with suitable carrier
molecules to
facilitate delivery to a target site (e.g., by complexing the molecules with
liposomes) and/or can
be bound to a targeting molecule (e.g., a ligand specific for a receptor
expressed on the surface of
a host cell infected by the microbe). Preferably, the targeting molecule
includes an intracellular
localization signal for delivering the antisense molecule to the interior of
the cell.
Therapeutic Peptides
As discussed above, candidate sequences, or sequences identified by these
sequences, can
be used to generate peptides. In some aspects, the peptides are administered
to the host organism
in an amount sufficient to enable the host organism to mount a protective
immune response
against the microbe. In a preferred embodiment, the peptides are used as a
vaccine.
Alternatively, or additionally, nucleic acid sequences which encode these
peptides and which are
operably linked to one or more promoter elements can be administered to the
host organism in an
amount sufficient to enable the host organism to mount a protective immune
response against the
microbe (e.g., providing a DNA vaccine). A protective immune response can
include the
production of macrophages which specifically recognize the microbe (e.g.,
during an
extracellular portion of its life cycle) and/or the production of cells which
produce neutralizing
antibodies which specifically bind to the microbe and which prevent the
microbe from infecting
further cells.
In some aspects, a plurality of peptides from the same microbe or a nucleic
acid
expressing the plurality of peptides is administered to the organism. In some
embodiments, the
microbe is isolated and nucleic acids removed, and the microbe itself is
administered to an
organism to generate a protective immune response (see, e.g., as described in
U.S. Patent No.
5,698, 430, the entirety of which is incorporated by reference herein).
23
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
Examples
The invention will now be further illustrated with reference to the following
examples. It
will be appreciated that what follows is by way of example only and that
modifications to detail
may be made while still falling within the scope of the invention.
Example 1
Computational subtraction was used to identify sequences in an EST library
(Unigene
library #271) from the HeLa cervical carcinoma cell line. This library
contains 7,073 EST's.
6,752 of these EST's comprise at least 100 discrete, unambiguous 15-mers
(e.g., sequences
whose nucleotide identity can be assigned at greater than 95% confidence
levels or 0% N's). A
system 1 according to the invention was used to compare the sequences in the
EST library
against known human mIRNA sequences, human repeat sequences, human
mitochondrial
sequences, the Human Genome Project (HGP) and Cetera Genomics Human Genomic
DNA
sequences and to eliminate matching sequences. Matches within mouse genomic
DNA
sequences (Cetera) were also searched for and removed under the assumption
that these would
represent unsequenced regions of the human genome.
Using a BLAST score cut-off of 60, equivalent to 30 consecutive identical
nucleotides,
and an expectation value of 10'' as a cutoff, the 7,073 EST's were pared down
to 144 non-
matching sequences. Application of a quality filter to set a minimum length
cut-off of 150
nucleotides and a 2.5% maximum ratio of ambiguous nucleotides (e.g., > 2.5%
N's) to non-
ambiguous nucleotides, decreased the number of matching sequences to 43. When
the 43
remaining sequences were matched to the GenBank nt database, 17 sequences
matched
additional human mIRNA sequences and 6 matched known E. coli sequences.
Thus, using the system l, 7,073 EST's in HeLa cells were reduced to a total of
22
sequences that failed to match human, mouse, or E. coli genomic sequences. Two
of these
sequences were subsequently determined to be identical to human papiltomavirus
(HPV) type 18
sequences. HPV is a cause of cervical cancer and HPV nucleic acids are known
to be present in
the HeLa cell line (see, e.g., Boshart et al., 1984, EMBO 3 5 : 1151-1157).
Two other HPV
sequences were present in the HeLa cell EST library, but were filtered out by
the system 1
because of a match to sequences in the Cetera genome database. These two
sequences match a
HPV type 45 sequence from the NCBI database that was included in the Cetera
genome
assembly but not in the NCBI assembly, thus, verifying the ability of the
system 1 to identify
microbial sequences through computational subtraction.
24
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
To determine which of the unmatched sequences (i.e., candidate sequences)
represented
unsequenced human genomic DNA and to determine whether candidate sequences
could be used
to identify pathogenic DNA (e.g., such as HPV DNA), PCR primers corresponding
to each of the
22 non-matching sequences were generated and tested on a panel of normal human
genomic
DNA samples and Hela cell genomic DNA. Ten primer sequences were capable of
amplifying
nucleic acids in all samples of human genomic DNA, while ten primer sequences
could not
amplify any samples, and two primers (corresponding to HPV sequences) were
able to amplify
only HeLa cell genomic DNA. The ten sequences which amplified all human
genomic DNA
samples are likely to represent previously unsequenced regions of the human
genome, while
those primer sequences unable to amplify sequences in any samples are likely
to represent
sequences brought together by splicing (and which are therefore too far apart
in genomic DNA to
be amplified), sequences of non-human origin, or sequencing errors.
These results demonstrate that the system 1 is capable of identifying
microbial sequences
(e.g., such as HPV sequences) by computational subtraction.
Example 2
Given the ability of computational subtraction to detect viral sequences in
HeLa cells,
computational subtraction was used to scan existing EST databases for
candidate microbial
sequences using the same method as described in example 1. EST's in a NCBI EST
database of
3,287,578 sequences were serially compared against filter databases using the
MEGABLAST
tool with a word-size of 24, to filter or subtract matching sequences. The
sequences were filtered
through a known human mRNA database (the NCBI Refseq human mRNA database),
which
after subtraction left 1,438,967 sequences, a human mitochondria database,
which after
subtraction left 1,409,118 sequences, a vector sequence database (the NCBI
UniVec database),
which after subtraction left 1,396,697 sequences, a human repetitive sequence
database (GIRST
HumRep), which after subtraction left 1,368,895 sequences, a human genome
database, which
after subtraction left 144,498 sequences, and a mouse genome database, which
after subtraction
left 137,011 sequences. To improve the sensitivity of the filtering process,
remaining EST's
were re-run against the filters at a lower word size (20) and matching EST's
were again
removed, leaving 120,792 EST's unmatched. The process was repeated using a
word size of 16,
leaving 102,009 EST's unmatched. This last sequence set was passed through a
quality filter as
described in Example l, to remove short (< 150 nucleotides) and ambiguous (>
2.5% N's)
sequences. At the end of this subtraction or filtering process, 65,839
sequences of 3,287,578
CA 02444604 2003-10-17
WO 01/54557 PCT/USOI/12736
EST sequences or 2% of the EST sequences in the NCBI EST database were found
not to match
a panel of human genomic or reference sequences.
Sequences were subsequently tested by BLASTN searches against GenBank nt
databases
(i.e., a database comprising multiple species' sequences, including microbial
sequences) using
(using a word size of 16) and by BLASTX searches against the nr non-redundant
protein
databases (using a word size of 3). A results database of these matches is
available at
http://www.hcs.harvard.edu/~weber/meyerson2/ nrnt.cgi, the entirety of which
is incorporated by
reference herein.
Despite filtering sequences against human genomic and other databases (e.g.,
removing
matching sequences), a significant fraction of the remaining EST's still
matched nucleotide
and/or protein sequences of known human origin. In total, 5,119 "non-matching"
candidate
sequences matched nucleotide sequences from Homo Sapiens using a BLASTN score
minimum
of 100, while the translations of 211 sequences, without nucleotide matches,
matched Homo
Sapiens protein sequences, with a BLASTX minimum score of 100. These data are
consistent
with the as-yet incomplete sequencing of the human genome.
Strikingly, a significant number of sequences with matches to viral, fungal,
bacterial, and
plant sequences were found in the non-matching, i.e., candidate sequence set.
A culled set of
matching species sequences was generated by excluding all vertebrate, as well
as Escherichia,
Saccharomyces, Drosophila, and Caenorhabditis sequences that might represent
library
contamination. Using BLASTN and BLASTX minimum scores of 100, 1,055 sequences
were
found which matched nucleotide sequences from the culled species (i.e.,
sequences representing
likely contaminants) and 759 sequences were found which matched culled protein
sequences but
not nucleotide sequences. Matches to microbial sequences are described in
Tables 1 through 3.
26
WO 01/54557 CA 02444604 2003-io-17 pCT/USO1/12736
Table 1. Viral
Genomes Wit
1 Nucleotide
Similarity
to Filtered
Human EST Sequences
Viral Species EST Library Tissue Types Most Common
Count C,'ount* Library No.***
Hepatitis B 33 '? adult liver, hepatocellular3618 (adult
virus liver)
carcinoma
Human 10 1 fetal liver 168 (fetal
liver)
spumaretrovirus
Cytomegalovirus9 3 nervous system, breast,2915 (nervous
uterus system)
Human adenovirus8 6 lymph, ovary (2)**, 2222 (lymph)
colon,
2 head and neck, lung
Simian sarcoma 7 6 breast 3633 (breast)
virus
Human 7 3 cervix, placenta, 271 (cervical
uterus,
papillomavirus tumor carcinoma cell
(subtypes 16 line, i.e.,
and HeLa)
18)
Stealth virus 4 2 head and neck (2) 4582 (head
1 and
neck)
Kaposi's sarcoma3 2 nervous, head & neck 2836 (nervous
associated virus system)
(HHV-8)
Hepatitis C 2 1 bone marrow 4862 (bone
virus
marrow)
Epstein-Barr 1 1 lymph 5167 (lymph)
virus
(HHV-4)
* "Library Count"
reflects the
number of libraries
in which EST
matches to
a particular
virus
were found.
** The total
number of different
libraries of
a given tissue
type is indicated
in parentheses
(if
greater than
one).
*** Library
numbers are
based on UniGene
assignments.
EST sequences that passed all filters (e.g., remained present after
computational
subtraction against one or more databases) were compared to GenBank's nt
database (a database
representing multiple species) using the MEGABLAST algorithm. Alignments with
a bit score
of 100 or greater were categorized as "matching" those in the nt database.
Sequences remaining
after subtraction which match viral genome sequences are shown in Table 1.
Included in these
sequences were sequences belonging to a variety of pathogenic viruses. As
shown in Table 1,
the most common viral match was to Hepatitis B virus sequences, for which
there were 33 EST
27
WO 01/54557 CA 02444604 2003-10-17 pCT/USO1/12736
matches in the databases. Thirty-two of these matches were derived from the
library GKC which
is made from normal liver tissue from a Chinese patient with hepatocellular
carcinoma.
Hepatitis B virus sequences are abundant in this library, representing 0.2% of
the 16,743 total
sequences in this library. As seen in Table 1, a variety of other pathogenic
virus sequences
including human papillomavirus; adenovirus; and a variety of herpesviruses,
including
cytomegalovirus, Epstein-Barr virus, and Kaposi's sarcoma herpesvirus; were
identified by
computational subtraction methods according to the invention.
Table 2, below, summarizes sequences remaining after computational subtraction
which
match bacterial sequences. After identifying expressed sequences as candidate
sequences not
found in the human genome, these sequences were compared to the GenBank nt
database using
the BLASTX algorithm (BLAST 2.0) and alignments with a bit score of 100 or
greater where
categorized as matches. Table 2 shows the ten most frequently appearing
bacterial sequences
after computational subtraction. As can be seen from Table 2, there are
numerous matches to
Pseudomonas aeruginosa sequences, a common pathogen as well as a commensal
organism. In
addition, there are numerous matches to other Pseudomonas species.
Table 2. Bacterial
Genomes With
Nucleotide Similarity
To Filtered Human
EST
Sequences
Bacterial speciesEST LibraryTissue Types Most Common
Count Count* Library No.***
Pseudomonas 304 85 breast (21)**, head 3025 (bone
and neck
aeruginosa (11), bone marrow (8) marrow)
Xylella fastidiosa92 32 head and neck ( 12), 1304 (breast)
breast
(10), stomach (3)
Pseudomonas sp. 56 11 breast (3), lymph 92),4873 (uterus)
B-
cells, muscle, ovary
Pseudomonas putida32 17 head and neck (6), 1148 (breast)
breast (5),
bone marrow (2)
28
WO 01/$4557 CA 02444604 2003-10-17 pCT/USO1/12736
Table 2 (cont'd).
Bacterial Genomes
With Nucleotide
Similarity To
Filtered Human
EST
Sequences
Caulobacter cresentus29 7 thymus (2), colon, 3587 (thymus)
lymph,
uterus
Mesorhizobium 26 13 lymph (2), thymus (2),2223 (lymph)
loti
foreskin, breast
Fusobacterium 17 5 head and neck (4), 4796 (head
uterus and
naviforme neck)
Leptotrichia-like17 10 uterus (9), head and 4796 (uterus)
sp. neck ( 1 )
* "Library Count"
reflects the
number of libraries
in which EST
matches to a
particular bacteria
were found.
** The total number
of different
libraries of
a given tissue
type is indicated
in parentheses
(if
greater than one).
*** Library numbers
are based on
UniGene assignments.
The more interesting category of bacterial matches is shown in Table 3 which
shows the
set of bacterial sequences whose conceptual translations match known bacterial
proteins and
which do not share significant nucleotide sequence similarity with known
bacterial nucleotide
sequences. These sequences were identified by passing EST sequences through
the filter
databases described above and comparing remaining sequences to the GenBank nt
database
using the BLASTN algorithm (with a threshold of 60 bits) and to the non-
redundant ("nr")
protein database using the BLASTX algorithm (setting a threshold of 100 bits).
EST's matching
the nr database but not the nt database were categorized as "translation-only
alignments." These
series of operations revealed numerous pathogens with matches only to
translated sequences.
Again, many matches were found to Pseudomonas aeruginosa sequences. Other
candidate
sequences included those whose translated sequences matched proteins of
Mycobacterium
tuberculosis, Vibrio cholerae and Neisseria meningitidis. This suggests the
presence of clones
representing novel unsequenced bacteria, highly related to these pathogens,
but previously
undescribed, in the libraries.
29
CA 02444604 2003-10-17
WO 01/54557 PCT/USO1/12736
Table 3. Sequences
in Human
EST Libraries
With Translation
Matches to
Bacterial
Sequences
Bacterial EST count Library Count**Most Common Most Common
Species* Tissue Types Library No.****
Pseudomonas 239 96 breast (20)***, 3587 (thymus)
head
aeruginosa and neck ( 14),
bone
marrow (8), CNS
(7)
Caulobacter 32 17 thymus (3), breast2223 (lymph)
crescentus (3), lymph (2)
Xylella fastidiosa22 14 lymph (3), thymus3587 (thymus)
(2), lung (2)
Mycobacterium14 8 thymus (2) 2223 (lymph)
tuberculosis
Streptomyces 29 14 thymus (2), breast2223 (lymph)
coelicolor (2), lymph (2)
Vibrio cholerae13 11 ovary (2) 2217 (B cells)
Bacillus subtilus14 10 heart (2) 47 (heart)
Neisseria 12 10 thymus (3), breast650 (pooled)
(2)
meningitidis
Pseudomonas 11 4 bone marrow (2),3587 (thymus)
putida lymph ( 1 ),
thymus
(1)
* Bacterial
species to
which candidate
sequences
are related,
as determined
by matching
conceptual
protein translations
of candidate
sequences
but not matching
nucleotide
sequences
of
candidate
sequences.
** "Library
Count" reflects
the number
of libraries
in which
EST matches
to a particular
bacteria
were found.
*** The total
number of
different
libraries
of a given
tissue type
is indicated
in parentheses
(if
greater than
one).
**** Library
numbers are
based on
UniGene assignments.
In each of the Examples discussed above, sequence analysis was performed using
sequence available from the NCBI (http://ncbi.nlm.nih.gov.), Celera Genomics
(http://www.celera.com) and the Genetic Information Research Institute (GIRD
(http://www.girinst.org). Human EST sequences and library information, Human
Genome
Project Sequences (phases 0-3), the Refseq human mRNA set, and UniVec vector
sequences
were downloaded from NCBI on March 6, 2001. The "nr" and "nt" BLAST databases
were
downloaded from NCBI on March 30, 2001. The human mitochondria) genome
sequence is
W~ 01/54557 CA 02444604 2003-10-17 pCT~S01/12736
GenBank accession # NC 001807. The HeLa cell EST library analyzed is available
as Library
271 (Stratagene HeLa cell s3 S 37216) in the UniGene resource at the NCBI web-
site. The
Celera draft of the human genom~ and the 3x coverage of shotgun sequence from
the mouse
genome were downloaded from (:elera's website in January, 2001. RepBase6.2 was
downloaded from the GIRI database on March 7, 2001.
Variations, modifications, and other implementations of what is described
herein will
occur to those of ordinary skill in the art without departing from the spirit
and scope of the
invention.
What is claimed is:
31
