![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
|
LETTER 15,000 Unique Zebrafish EST Clusters and Their Future Use in Microarray for Profiling Gene Expression Patterns During Embryogenesis1Functional Genomics Lab, Institute of Molecular and Cell Biology, Singapore 117609; 2Molecular and Developmental Immunology Lab, Institute of Molecular and Cell Biology, Singapore 117609
A total of 15,590 unique zebrafish EST clusters from two cDNA libraries have been identified. Most significantly, only 22% (3437) of the 15,590 unique clusters matched 2805 (of 15,200) clusters in the Danio rerio UniGene database, indicating that our EST set is complementary to the existing ESTs in the public database and will be invaluable in assisting the annotation of genes based on the upcoming zebrafish genome sequence. Blast search showed that 7824 of our unique clusters matched 6710 known or predicted proteins in the nonredundant database. A cDNA microarray representing  3100 unique zebrafish cDNA
clusters has been generated and used to profile the gene expression
patterns across six different embryonic stages (cleavage, blastula,
gastrula, segmentation, pharyngula, and hatching). Analysis of
expression data using K-means clustering revealed that genes coding for
muscle-specific proteins displayed similar expression patterns,
confirming that the coordinate gene expression is important for
myogenesis. Our results demonstrate that the combination of microarray
technology with the zebrafish model system can provide useful
information on how genes are coordinated in a genetic network to
control zebrafish embryogenesis and can help to identify novel genes
that are important for organogenesis. [Supplemental material is available online at www.genome.org. The sequence data from this study have been submitted to EMBL under accession nos. AL901610–AL928536.]
In recent years, zebrafish has been adopted as a model system for studies of vertebrate development because of its unique features favorable for genetic studies compared with other vertebrate systems. These features include a reasonably short lifetime, a large number of progenies, external fertilization and embryonic development, and translucent embryos (Talbot and Hopkins 2000  ). The relative ease of obtaining haploid and homozygous diploid
individuals by gynogenesis offers another advantage for genetic
analysis (Streisinger et al. 1981, 1986). Thus far, genetic linkage
maps with a high density of markers covering the whole genome are
readily available in the database (Knapik et al. 1998; Postlethwait et
al. 1998; Gates et al. 1999; Shimoda et al. 1999). Zebrafish has also
been regarded as an excellent model system for the studies of human
disease because many zebrafish mutants with phenotypes such as
disorders in hematopoiesis, cardiovascular generation, and kidney
development are reminiscent of human disease states (for review, see
Dooley and Zon 2000). The upcoming completion of sequencing of the
zebrafish genome will no doubt facilitate our genetics and genomics
studies in zebrafish in the future.
Two large-scale screens using chemical mutagens have been performed and
>1500 phenotypic mutants corresponding to over 500 genes have been
generated to identify genes required for the early developmental
processes in zebrafish (Driever et al. 1996
Although many genes important for cell signaling or organogenesis have
been studied in zebrafish, global analysis of gene expression will
likely provide more information regarding how gene regulation is
coordinated in this genetic network. In addition, systematical analysis
of the differences in the patterns of gene expression among different
stages or between a mutant and the wild-type control will also provide
an opportunity to identify those functional but unknown genes that
might be important for controlling embryogenesis and organogenesis. The
ease of harvesting a large amount of near-synchronized embryos provides
an excellent opportunity to obtain mRNA from embryos from different
embryonic stages for the study of gene expression patterns. Here we
report that a total of 26,927 ESTs (length >200 bp, average length
473 ± 95 bp) have been obtained from two zebrafish cDNA libraries,
one a primary library (Z1) and the other a normalized library (Z2). We
also report the result from clustering analysis of these ESTs and
summarize the result from blast search against a public database. To
have a global view of gene expression profiles during zebrafish
embryogenesis, we generated a cDNA microarray carrying
ESTs From the Z1 Library Total RNA was extracted from zebrafish at different growth stages (for details, see Methods; Kimmel et al. 1995 ). The unamplified primary
double-stranded cDNA (ds-cDNA) derived from the mRNA sample was used
for the construction of our Z1 cDNA library. Plasmid DNA from an
individual clone randomly picked from the Z1 library was subjected to
EST sequencing. As shown in Table 1, 2556
unique cDNA clusters were identified from a total of 11,908 individual
clones with length >200 bp. Overall redundancy is close to 78%
(unique clusters against total sequence runs, the same following),
which is normally seen for nonnormalized cDNA libraries (statistics
from ftp://ftp.ncbi.nlm.nih.gov/repository/Unigene/Dr.seq.uniq).
Statistical analysis showed that only 1660/2556 clusters are each
represented by a single clone and the rest are each represented by >2
clones (Table 2). Notably, 16 genes,
accounting only for 6.3% (16/2556) of unique cDNA clusters, are each
represented by >100 clones (Table 2) and contribute significantly to
the high redundancy (3638/11908 = 31%). Eleven of these 16 genes
share homology with genes of known/putative function in the
nonredundant (nr) database, and their protein products are involved in
different cellular activities (Supplementary Table 1). The cluster for
the Y-box binding protein gene (also known as nuclease-sensitive
element binding protein; Didier et al. 1988; Gai et al. 1992; Grant and
Deeley 1993) contains 648 individual clones, indicating a high
abundance of the transcript of this gene in our mRNA sample
(Supplementary Table 1). Five of the 16 high-abundance genes do not
match any known gene or predicted protein (Supplementary Table 1).
Because the majority of our ESTs in Z1 contain a poly(A) tail and
because sequencing was done from the 3' end, we cannot exclude the
possibility that the sequences obtained for these five genes may not
include the ORF if they have a long 3'-untranslated region (3'-UTR)
sequence.
ESTs From the Normalized cDNA Library (Z2) To reduce such high redundancy during EST sequencing and to speed up the identification of unique cDNAs, we constructed a normalized cDNA library (Z2), as illustrated in Figure 1 (for details, see Methods). The strategy used to construct our normalized cDNA library was based on the principle that, during cDNA annealing, rare cDNA transcripts anneal less rapidly than do abundant cDNA species; thus, the single-stranded fraction of cDNA (ss-cDNA) becomes progressively more normalized during the course of annealing (Ko 1990 ; Patanjali et al. 1991). To determine the normalization
efficiency, we hybridized 10,000 clones of each library constructed
from the prenormalized and the normalized cDNA, respectively, with
probes derived from the  -tubulin gene (Fig.
2C). A dramatic reduction in the number of
positive clones was observed in the normalized cDNA library (8;
NIII) when compared with the prenormalized library (67; SI). On the
other hand, the number of clones for the relative low-expressing gene
RAG1 (recombination activating gene 1) was increased from zero
(in SI) to four (in NIII). Clones from this high-quality library (Z2)
were used in our new EST sequencing project. As shown in Table 1,
13,308 unique cDNA clusters were obtained from a total of 15,019 ESTs
of length >200 bp. Overall redundancy in the Z2 library is only
12% and this rate is significantly lower than that in the Z1
library (Table 1). Statistical analysis showed that 11851/13308
clusters are each represented by a single clone and the rest are each
represented by more than two clones (Table 2). The cluster for the
beta-lactamase gene contains seven individual clones and is most
abundant in the Z2 library, which significantly contrasts with the
large number of redundant clones in the Z1 library (Table 2), further
demonstrating the effectiveness of normalization.
Statistical Analysis of ESTs From Z1 and Z2 Libraries Combining Z1 and Z2 libraries, a total of 15,590 of unique cDNA clusters were identified (Table 1; Supplementary Table 2). Large-scale EST sequencing for zebrafish has also being carried out at Washington University (WashU) headed by Dr. S. Johnson, and the total EST collection is approaching 200,000 (http://zfish.wustl.edu/). A portion of these EST sequences was subjected to clustering analysis and the result was deposited in the UniGene database in NCBI (ftp://ftp.ncbi.nlm.nih.gov/repository/Unigene/Dr.seq.uniq). As released on June 14, 2002, a total of 15,200 clusters were identified from a total of 188,259 ESTs and other sequences ( 90% redundancy).
Surprisingly, only 22% of our unique clusters (3437/15590) matched
2805/15,200 D. rerio UniGene clusters (Table 1, Supplementary
Table 3) in this UniGene database. This ratio is much lower than that
obtained from comparison between the EST sequences from the zebrafish
embryonic inner ear (44%) and the UniGene database (Coimbra et al.
2002). Thus, it is reasonable to estimate that, combining our and
WashU's EST set, 27,000 unique clusters have been identified and
these ESTs will be invaluable in assisting with the annotation of the
zebrafish genome in the near future. Because many cDNAs contain more
than one EcoRI site and because EcoRI fragments were
used to construct the normalized cDNA library, the final number of
unique clusters is likely slightly overestimated. This is reflected in
the fact that 3437 of our EST clusters matched only 2805 Unigene
clusters in the public database.
The translations in the six phases of all 15,590 unique clusters in our
database were compared with nr (released on 13 May 2002;
ftp://ftp.ncbi.nlm.nih.gov/blast /db/nr). As shown in Table 1,
7824/15590 unique clusters in total have hits
(P<10–8) in nr and these hits represent 6710
distinctive genes that include 5180 known genes either corresponding to
D. rerio known genes or to known genes with a wide range of
functions in other organisms (Supplementary Table 4). As
expected, genes encoding for metabolic enzymes (including hydroxylase,
oxidase, reductase, dehydrogenase, synthase, metabolic kinase and
phosphatase, transferase, and protein degradation enzymes) are most
abundant in our EST set, and in total 1058/6710 (
Generation of Zebrafish cDNA Microarray PCR products generated from 11,480 individual clones from the Z1 library were arrayed on glass slides, and, at the same time, these clones were subjected to DNA sequencing. In total, sequences of 10,518 EST clones were obtained (Supplementary Sequence File: in FASTA format) and the alignment result revealed that this set of ESTs represents 3100 unique cDNA clusters (Supplementary Table 5). Blast search using six frames of all ESTs revealed that 4519 ESTs had hits (P < 10–5) in the nr protein database (ftp://ftp.ncbi.nlm.nih.gov/blast/db/nr) and these 4519 hits represent 728 distinct proteins (Supplementary Table 6). Both muscle-specific proteins (Table 4) and ribosomal proteins (Supplementary Table 7) are well represented in this EST set. The remaining 5999 did not have a hit in the database. Nevertheless, for reasons mentioned earlier, the number of sequences on the array with no hits in our blast search is likely to be overestimated.
Gene Expression Profiling During Zebrafish Embryogenesis Zebrafish embryogenesis occurs during the first 3 d after fertilization and this period can be broadly categorized into seven stages: zygote, cleavage, blastula, gastrula, segmentation, pharyngula, and hatching (Kimmel et al. 1995 ). By the end of the third day, most of
the major organ systems in the organism, including brain, nervous
system, muscle, heart, blood, gut, liver, eyes, and ears have already
been initiated and some are even properly formed from the totipotent
fertilized egg (Kimmel et al. 1995). Under optimal growth conditions,
the time span between each sequential stage is almost invariable, and
the accomplishment of embryogenesis is the result of precise
coordination of the unidirectional transitions between each stage under
the control of a precise genetic network.
Total RNA was extracted, respectively, from unfertilized eggs (E0) and
near-synchronized embryos at the stage of cleavage (E2), blastula (E3),
gastrula (E4), segmentation (E5), pharyngula (E6), and hatching (E7)
for mRNA preparations. Because it represents the maternal transcripts
of the basal level from where de novo embryonic gene expression at
different stages can easily be compared, the unfertilized egg (E0) was
used as a reference sample against which samples from other stages were
compared. In each case, the experiment was repeated, reversing the
fluorescent dye. For each slide, the fluorescence intensities of the
two probes hybridized to each spot were normalized and the ratios were
then calculated (Supplementary Table 5). To identify those genes whose
expression is regulated during embryogenesis, we determined the number
of clones showing a ratio (against E0) greater or less than twofold
(log value >1 or <–1) for each stage (Table
5). For example, at the E3 stage, 356 of
3100 total clusters examined showed a significant increase (log ratio
To study the expression pattern of each individual gene during embryogenesis in detail, we analyzed normalized data (Supplementary Table 5) from 24 microarray experiments (4 experiments for each stage) using hierarchical clustering and K-means clustering methods (EPCLUST: http://ep.ebi.ac.uk/EP/EPCLUST/). Hierarchical clustering and K-means clustering grouped genes with similar expression patterns that displayed dynamic changes of expression. In general, expression of most genes on the array examined displayed one of the following expression patterns: up-regulated, down-regulated, and fluctuating (data not shown). We focused on groups displaying patterns of either up- or down-regulated. All expression data were clustered into 20 groups using the K-means clustering method (http://ep.ebi.ac.uk/EP/EPCLUST/). Seven groups were selected to form five different expression patterns (Fig. 3), including early up-regulated (49 cDNA clusters from E3 and E4 stages; Fig. 3A; Supplementary Fig. 1), middle up-regulated (350 cDNA clusters from E4 and E5 stages; Fig. 3B; Supplementary Fig. 2), late up-regulated (64 cDNA clusters from E6 and E7 stages; Fig. 3C; Supplementary Fig. 3), early down-regulated (151 cDNA clusters from E3 and E4 stages; Fig. 3D; Supplementary Fig. 4), and late down-regulated (251 cDNA clusters from E4 and E5 stages; Fig. 3E; Supplementary Fig. 5). The early down-regulated group contains genes encoding for proteins such as claudin-like protein ZF-A89 and ZF-A9 (Fig. 3D; Supplementary Fig. 4). The late down-regulating subgroup includes genes encoding for proteins such as the p32 subunit of splicing factor SF2, single-stranded D-box binding factor 2, and GAP-associated phosphoprotein p62 (Fig. 3E; Supplementary Fig. 5). As ribosomal genes (Supplementary Table 7) and muscle genes are well represented in this EST set (Table 4), we examined their expression patterns from the E2 to E7 stages. The expression of most ribosomal genes belong to the middle up-regulated genes, indicating the initiation of active protein biosynthesis at this stage (Fig. 3B; Supplementary Fig. 2), whereas a significant number of muscle genes fall in the late up-regulated group (Fig. 3C; Supplementary Fig. 3), indicating the maturation of myogenesis at this stage. The coordinate actions displayed by muscle genes and ribosomal genes indicate that coordinate regulation of gene expression is not only important for organogenesis but also for subcellular organogenesis.
Coordinate Expression of Muscle Genes The abundance of muscle genes on our microarray provided us with a chance to study their coordinate actions and myogenesis. In zebrafish, myogenesis starts at the segmentation stage and strong evidence has shown that myogenic transcriptional factor MyoD (Weinberg et al. 1996 ),
Engrailed (Halpern et al. 1993), and Hedgehog (Roy et al. 2001) are
involved in controlling muscle initiation and differentiation.
Zebrafish embryonic muscle contains three distinct fiber types: muscle
pioneer slow muscle, nonpioneer slow muscle fibers, and fast muscle
fibers (Barresi et al. 2000). Ten zebrafish muscle-specific genes have
been identified and analyzed previously (Xu et al. 2000) and the
reported result strongly indicated that the expression of these genes
is coordinately regulated. Blast search revealed that 27 unique
clusters on our array are likely genes encoding different muscle
proteins, including skeletal  -actin, myosin light chain, troponin,
tropomyosin, and other regulatory proteins (vimentin, desmin, etc.).
Consistent with a previous report (Xu et al. 2000), hierarchical and
K-means clustering showed that many known muscle genes in our data set
displayed a similar expression pattern (Fig. 3C; Supplementary Fig. 3).
The expression patterns of six well-known muscle genes from microarray
hybridization were shown as examples in Figure
4. Their expression
patterns from E2 to E7 stages were further confirmed via RNA gel blot
hybridization and their identities as muscle-specific genes were
verified via in situ hybridization (Fig. 4).
Prediction of Putative Unknown Muscle Genes Using Support Vector Machine (SVM) The SVM method is mainly used to predict putative functional unknown genes that might function in the same pathway or be coregulated in the same organ or tissue or subcellular organelle as are those known genes based on their expression patterns. SVM uses a training set to specify in advance which data should cluster together (Brown et al. 2000 ). This
method is mainly based on the assumption that genes of similar function
display similar expression patterns (Eisen et al. 1998). A set of known
ESTs (2500) was used as a training set to allow the SVM to learn to
discriminate between muscle and nonmuscle genes (Table
6). To examine the quality of training, we
tested the trained SVM on another set of known ESTs (1387). As shown in
Table 6, the majority of predicted positives are true positives
(44/54), whereas the majority of predicted negatives are true negatives
(1286/1333), proving that the training was fairly successful. When the
expression data of the remaining unknown ESTs (7562) were subjected to
SVM analysis, 110 candidates, corresponding to 56 unique cDNA clusters
and 30 unsequenced cDNAs, were predicted as true positives (Table 6;
Supplementary Fig. 6). One such cDNA (clone ID 160-D11) was chosen for
further studies because a preliminary blast search using the available
sequence of 160-D11 did not identify any hit in GenBank.
A probe derived from 160-D11 was used for RNA gel blot hybridization and a transcript of size 6.7 kb was identified. Examination of transcript levels of 160-D11 at different stages (E2–E7) showed that its expression pattern changed in a similar way as did other muscle-specific genes (Fig. 4), which is consistent with the pattern obtained from microarray hybridization. Furthermore, in situ hybridization showed that 160-D11 was specifically expressed in muscle (Fig. 4). Because of the large size of transcript (6.7 kb) for 160-D11 as revealed by Northern analysis, repeated attempts to get the full-length sequence for 160-D11 through 5' racing were unsuccessful. Nevertheless, we managed to get 2.5 kb of 160-D11 cDNA sequence and blast search did not identify any homology with any known gene in the database. Based on the fact that the size of the 160-D11 transcript is 6.7 kb, we could not exclude the possibility that the sequence
obtained does not contain the ORF, and thus we could not fully exclude
the possibility that 160-D11 is a known gene.
In this report, we described the EST sequencing results from Z1 and Z2 libraries. Clustering analysis showed that clone redundancy in Z1 is much higher than in Z2. Apparently, the low redundancy in Z2 resulted from successful normalization of cDNA derived from our mRNA sample. As indicated from screening 10,000 colonies, the number of beta-tubulin and RAG1 clones was down approximately eightfold and up approximately fourfold, respectively. Presumably, after normalization, most genes would be more or less evenly represented in the cDNA pool used for library construction. If we could propose that there were 30,000–40,000 expressed genes represented in our RNA sample collection, mathematically, randomly picking 20,000 clones from this normalized library (we picked 200 x 96 clones) would avoid the problem caused by random cloning and likely provide us with a high chance of obtaining the clones representing unique ESTs, although, inevitably, some clones were redundant. We believe this is the case; as proof, when compared with the public Unigene database, 3437 of our unique EST clusters matched 2805 of the Unigene cluster. The Unigene cluster normally covers a significant length of the corresponding cDNA. Had the problem of random cloning been a contributing factor, a higher number (at least 2 x 2805) of our ESTs would have been expected. In addition, as further evidence, when all unique ESTs were used to blast the nr database, we got 7824 hits representing 6710 distinctive genes. As a significant number of genes are deposited as full length in nr, had the random cloning caused a problem, many fewer distinctive genes would be identified. Considering cost and efficiency, the high quality of the normalized cDNA library (Z2) makes it ideal for EST sequencing. The lower rate of overlapping with the Unigene set in the public database, the wide-range coverage of genes involved in diverse cellular activities, and the large number of novel sequences in our EST database provide us with a rich resource to generate the zebrafish cDNA microarray. Because all individual clones have been stored carefully, EST clones could be accurately retrieved and provided on the basis of microarray results. We are currently generating a zebrafish cDNA microarray using ESTs representing our unique cDNA clusters to provide a public service to the fish community worldwide.
The accomplishment of embryogenesis is marked by the formation of
almost all necessary organs and tissues for a living embryo. This
process is under the control of a genetic network that relies on the
precise coordinate actions of many genes that are essential for normal
development. For decades, great efforts and also great achievements
have been made to identify genes important for organogenesis and to
study how these genes interact with one another to control
embryogenesis. Because of some of its unique features favorable for
genetics studies, zebrafish has recently been chosen as the model
system to study vertebrate development (Talbot and Hopkins 2000
Construction of a Primary cDNA Library (Z1) and a Normalized cDNA Library (Z2) Total RNA was extracted from zebrafish (local wild type) at different stages (for Z1, stages used were 12-h-, 2-d-, 4-d-, 1-wk-, 2-wk-, 1-mo-, 2-mo-, and 3-mo-postfertilization, respectively; for Z2, stages used were 4-h-, 2-d-, 4-d-, 1-wk-, 2-wk-, 1-mo-, 2-mo-, and 3-mo- postfertilization, respectively; Kimmel et al. 1995 ), using the
Tri Reagent according to the manufacturer's protocol (Molecular
Research Centre). An equal amount of total RNA from each of the stages
was mixed for mRNA purification using an mRNA purification kit
(Promega). For the Z1 library, both the first- and second-stranded cDNA
was synthesized using a cDNA synthesis kit (GibcoBRL; primer for
first-strand cDNA:
5'-GAGAGAGAGAGAGA GAGAGAACTAGATCTCGAGTTTTTTTTTTTTTTTTTT; adapters are
composed of 10- and 14-mer oligonucleotides, which are complementary to
each other with EcoRI cohesive end: 5'-OH-AATTCGGCACGAGG-3'
and 3'-GCCGTGCTCC-P-5'). A size selection was performed for cDNAs of
0.5–2.0 kb before cloning them unidirectionally into predigested
vector (5'-EcoRI and 3'-XhoI sites; ZAP
Express cDNA Synthesis kit, Stratagene). Transformed cells were plated
at 3000 cfu per 24 x 24 cm2 Qtray (Genetix), and grown
overnight. An individual colony was selected and placed (Qpix from
Genetix) in 96-well deep well plates with 1 mL
LB/ampicillin per well. The overnight cultures were divided into three
portions (Biomek FX, Beckman): (1) 100 µL for PCR, (2) 50 µL for
glycerol stock, and (3) the remainder for minipreps.
For the Z2 library, oligo(dT)20-V (V = G, C, A) was used as
primer for the first-strand cDNA synthesis and the second strand was
synthesized as described (GibcoBRL). Two primers, LLR1A
(5'-gagatattagaattctactc-3') and LLR1B (complementary strand
5'-gagtagaattctaatat-3'; Ko 1990
High-Throughput Sequencing and Sequence Assembling
Sequence Comparison Against Public Databanks
Microarray Preparation
Array Printing
Controls
Fluorescent Probe Preparation
mRNA and Fluorescent Probe Preparation
Array Hybridization
Data Processing and Analysis
Northern Blot Analysis
In Situ Hybridization
ftp://ftp.ncbi.nlm.nih.gov/blast/db/nr; Non-redundant (NCBI). ftp://ftp.ncbi.nlm.nih.gov/repository/Unigene/Dr.seq.uniq; National Centre for Biotechnology Information (NCBI). http://ep.ebi.ac.uk/EP/EPCLUST/; Expression Profiler of European Bioinformatics Institute (EBI) for hierarchy clustering and K-means clustering.
http://theoval.sys.uea.ac.uk/ http://www.mrc-lmb.cam.ac.uk/pubseq/manual/gap4_unix_toc.html; MRC gap4 program in Staden Package for sequence alignment. http://www.mrc-lmb.cam.ac.uk/pubseq/manual/pregap4_unix_toc.html; MRC pregap4 program in Staden Package for sequence editing. http://www.tigr.org/software/assembler; Tigr-Assembler. http://zfish.wustl.edu/; Washington University–Zebrafish Genome Resources Project
We thank Dr. Steve Johnson for his useful discussion on EST sequencing during his visit in Singapore. We thank Dr. Robert Schaffer, Matthew Larson, and Dr. Ellen Wisman (Michigan State University) for helpful advice on techniques in microarray. We thank Dr. Wai Ming Kong and Dr. Keng Wah Choo (Nanyang Polytechnic, Singapore) for helping with data normalization. We thank Dave Oh in the IMA sequencing unit for performing part of the EST sequencing. This work is financially supported by the Agency for Science, Technology and Research (A*STAR) in Singapore. The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.
3 These authors contributed equally to this work.  E-MAIL pengjr{at}imcb.a-star.edu.sg; FAX 65-68727007. E-MAIL zilong{at}imcb.a-star.edu.sg; FAX 65-68727007. Article and publication are at http://www.genome.org/cgi/doi/10.1101/gr.885403.
|
Top
ABSTRACT
RESULTS
1) and 160 showed a significant decrease (log ratio 
–1) in their
expression levels, indicating that activities at this stage might not
need the de novo expression of many genes (Table 
h); (E4) gastrula (5
to 5
h); (E5)
segmentation (6 to 10 somites); (E6) pharyngula (24 h); (E7) hatching
(72 h). M. thermautotrophicus = Methanothermobacter
thermautotrophicus.
DNA) were placed on our array. They were each
spotted on six subgrids of an array, at the four corners of each (i.e.,
six spots per control across the whole array). These control inserts
were amplified using the universal primer pairs on their respective
vectors, followed by similar precipitation and resuspension in spotting
buffer as described earlier. Of the 10 negative controls, the first 4
were also used as spiking controls for data normalization.