Info

Automatic peak detection and calibration:

i- MS intensity data for each MS spectrum at position (Xj, yj

Set of identification data: protein ID as * function of position

Peptldent Smartldent

Identification In SWISS-PROT / TrEMBL

D) Visualisation of results Automated data processing and identification C)

Schematic representation of the molecular scanner process.

shots on the sample plate and the naming of the MS files is controlled by dedicated software.

C) Processing of the MS data and protein identification. A flexible and interactive tool has been developed to automatically perform peak detection and calibration on all spectra. The positions on the sample plate are converted to apparent molecular mass (Mr) and pi values. The PMF data of all spectra, together with the calculated pi and Mr and other user defined parameters (such as mass tolerance, considered chemical modifications, species, etc.), are automatically submitted to Peptldent for identification

D) Analysis of the results: creation of virtual maps. The identification results of Peptldent are represented as an annotated image that can be read by the Melanie image analysis software. The data process program generates a first virtual "2-D-map", i.e. a 3-D image where the x and y co-ordinates relate to pi and Mr values, respectively. The z values reflect in gray scales the intensity of the MS spectra, as defined by the sum of the intensities of the MS signals in the considered MS spectra. The image also contains the identification results, which can be highlighted as labels in Melanie. in addition, this image includes, other protein attributes that reflect the number of dimensions of the image, i.e. pi, Mr,

IDH ECOLI

Interpretation of results obtained with the molecular scanner. 250 |g of E. coli were separated on a mini-2DE on a narrow range pI gradient (5-5.5). The gel was submitted to DPD (full in-gel digestion followed by a through trypsin digestion). 1.3 x 0.9 cm2 were scanned with a Voyager Elite MALDI-TOF MS. The protein identifications were performed with on-line

Interpretation of results obtained with the molecular scanner. 250 |g of E. coli were separated on a mini-2DE on a narrow range pI gradient (5-5.5). The gel was submitted to DPD (full in-gel digestion followed by a through trypsin digestion). 1.3 x 0.9 cm2 were scanned with a Voyager Elite MALDI-TOF MS. The protein identifications were performed with on-line calls to Peptldent. The area represents a pl range of about 5.1-5.25 and a Mr range of about 25-45 kDa. The grey scale intensity of this MS image represents the intensity of the MS spectra. The white dots show the positions at which the IDH_ECOLI protein was identified. Similarly, the position of PCK_ECOLI, 6PGD_ECOLI and METK_ECOLI are shown in circles.

identification labels (SWISS-PROT or TrEMBL AC numbers, ID labels), peptide masses, MS intensities as well as accesses to the detailed PeptI-dent annotations (number of missed cleavages, annotated modifications, chemical modifications of Cys and Met residues, peptide sequences).

The user can choose to filter part of the information and visualise only particular aspects of the multi-dimensional image. Proteins or peptides can be visualized by searching the positions at which they have been identified. The z intensity can represent the number of peptides found to match the queried protein in the identification process, for instance, or the sum of the MS intensities of the peptide masses matching the queried protein (for examples of data interpretation and visualisation see Figure 4.13).

Other techniques

The peptide mass fingerprinting identification method is a very accurate and quick approach, particularly using MALDI MS as an analytical tool. This method is optimal for the identification of proteins that are annotated in sequence databases. There are, however, a number of limitations to this technology. A number of small proteins are not producing a reasonable number of peptide ions to allow for unambiguous identification only by their masses. Very basic, very small (< 15 kDa) and highly hydrophobic proteins are difficult to separate with 2-DE technology. Other type of gel-based separation can help in these cases, such as Tris-Glycine gels or 1-D SDS gels. Due to the complexity of the mixture, it is often reasonable to pre-fractionate the biological sample of interest [69].

To complement the peptide mass fingerprinting approach, a variation of protein identification by mass spectrometry, called MS/MS, is often used. In this approach, the peptides generated after endoproteolytic cleavage of the proteins are separated in a mass spectrometer equipped with a fragmentation device. From all ions of the peptide mass fingerprinting spectra, selected ions are further fragmented. In a simplistic description, these fragments produce a set of masses in the so called MS/MS spectra, the mass differences of which correspond to masses of amino acid residues. The peptide sequence can therefore be deduced. In this approach, only a few spectra of good quality are needed to unambiguously identify a protein in a database. From all available machines capable of MS/MS, the ion-trap MS and the Q-TOF MS are those whose controlling software allows for the most automatic and data-dependent sample handling. Two interpretation approaches are derived from the MS/MS technique. The peptide-sequence-tag approach [8] requires the user to interpret the MS/MS spectrum, i.e. to ''read'' a sequence tag associated to the spectrum. The database is searched using the sequence tag and the mass of the entire ion. Yates and colleagues [58, 60] introduced a more automated approach. Their SEQUEST program compares uninterpreted MS/MS spectra with a library of predicted MS/MS spectra from a chosen database. Those two algorithmic approaches and others can be automated and linked to a launching system. SEQUEST is integrated in a package provided with the Finnigan Corporation (http:// www.finnigan.com) ion-trap MS instruments. A very similar approach is provided in the MASCOT software (Matrix Sciences Ltd, http://www. matrixsciences.com). The other programs can be reached through the ExPASy tools page (http://www.expasy.ch/tools). For a review on protein identification by mass spectrometry see Corthals et al. [5]. For a review on protein identification tools see Binz et al. [49].

Conclusion

Today's advances in technology allow us to analyze complex systems such as entire proteomes. Also, automation in proteomics generates a huge amount of heterogeneous data. This requires the integration of a Laboratory Information Management System (LIMS) connected to a full sample-tracking concept. Global solutions are still not commercially available for the pro-teomics field and do not cover a full system. A number of companies are currently developing such systems, like Applied Biosystems (Proteomics

Solution I and SQL LIMS) or Micromass-Bio-Rad (Proteome Works System), among others.

Comprehensive interpretation of the experimental proteomic data is only reachable in a step-by-step approach. This means that the experimentalist must have the possibility to come back to the original data at any time, in order to complete its characterization and in order to perform efficient comparison studies. In fact, as protein sequence databases are not complete, a number of identifications are failing, especially in eukaryotes. As the databases are regularly updated, identification must be regularly re-run with the incorporated data. This must regularly and automatically launch new calls to the identification tools and then update the results and the interpretation.

Moreover, even if all proteins are identified, there is the need to characterize them further. The importance of post-translational modifications in biological activity is proven, as well as that of amino-acid substitutions. These properties are often difficult to detect automatically, as they are often missing information in databases. They are, however, of potentially huge importance for the understanding of biological activity and of the presence of a diseased state.

It is now well established that combination of genomics and proteomics technologies will have a major impact in the future in: i) identifying gene products related to disease states that will further speed up the discovery of potential molecular markers and therapeutic targets, ii) understanding the mechanisms of action of drugs, including their pharmacological and toxic effects that may reveal much about the mechanisms of the disease, and iii) the discovery and design of new drugs that improve the diseased state.

In this chapter we have described a range of both wet-lab, experimental and dry-lab, bioinformatics approaches to extract biologically relevant information from one or several proteomes. In order to solve medical questions, such as "what is the molecular mechanism leading to a diseased state?" or "what kind of diagnostic test do I have to perform to decide for a diseased state?" or "what are the side effects of the absorption of a specific drug on the general state of a patient?", the presented methods might be appropriate. Very often, though, the question cannot be answered without combining proteomics with transcriptomics (i.e. the study of the RNAs, the transcription products of genes), genomics and more classical chemistry methods. In all these cases, the researcher often generates a huge amount of experimental data. These data, alone or in combination with other researchers' experiments, contain a limitless source of information. The huge variety of databases available on the web and the amount of information included are to be compared with the experimental data. There are big challenges for bioinformatics to handle very heterogeneous data formats, to correlate highly variable types of biological information, to extract their relevance for a particular question, to perform the appropriate analyses, and to provide results in a form that is valuable for high quality interpretation.

This is the technical challenge that will lead to discoveries in prognostics, diagnostics, and drug therapies.

References

1 Wilkins M. R., Williams K. L. (1997a) J. Theor. Biol. 186, 715.

2 Wilkins M. R., Williams K. L., Appel R. D., Hochsteassee D. F. (1997b) Proteome Research: New Frontiers in Functional Genomics, Springer-Verlag Berlin Heidelberg New York.

3 O'Faeeell P. H. (1975) J. Biol. Chem. 250, 4007-4021.

4 Klose H. (1975) Humangenetik 26, 231-238.

5 Coethals G. L., Gygi S. P., Aebeesold R., Patteeson S. D. (1999) in: Proteome research: 2D Gel Electrophoresis and Detection Methods, Ed. Rabilloud, Springer-Verlag New York.

6 Hoogiand C., Tonelia L., Sanchez J.-C., Binz P.-A., Baieoch A., Hochsteassee D. F., Appel R. D. (2000) Nucl. Acids Res. 28, 286-288.

7 Ceemona O., Muda M., Appel R. D., Feutigee S., Hughes G. J., Hochsteassee D. F., Geinoz A., Gabbiani G. (1995) Exp. Cell. Res. 217, 280-287.

8 Mann M., Wilm M. (1994) Anal. Chem 66, 4390-4399.

9 Hochsteassee D. F., Appel R. D., Vaegas R., Peeeiee R., Raviee F., Funk M., Pellegeini Ch., Mullee A. F., Scheeeee J.-R. (1991) MD-Computing 8, 85-91.

10 Bienvenut W. V., Sanchez J.-C., Kaemime A., Rouge V., Rose K., Binz P.-A., Hochsteassee D. F. (1999) Anal. Chem. 71, 4800-4807.

11 Binz P.-A., Müllee M., Walthee D., Bienvenut W. V., Geas R., Hoogland C., Bouchet G., Gasteigee E., Fabbeetti R., Gay S., Paiagi P., Wilkins M. R., Rouge V., Tonella L., Paesano S., Rosseliat G., Kaemime A., Baieoch A., Sanchez J.-C., Appel R. D., Hochsteassee D. F. (1999a) Anal. Chem. 71, 4981-4988.

12 Bjellqvist B., Righetti P. G., Gianazza E., Güeg A., Postel W., Westeemeiee R. (1982) J. Biochem. Biophys. Methods 6, 317-339.

14 Appel R. D., Paiagi P., Walthee D., Vaegas J. R., Sanchez J.-C., Pasquali C., Hochsteassee D. F. (1997) Electrophoresis 18, 2724-2734.

15 Andeeson N. L., Swanson M., Gieee F. A., Toliaksen S., Gemmel A., Nance S., Andeeson N. G. (1986) Electrophoresis, 7, 44-48.

16 Richaedson F. C., Steom S. C., Copple D. M., Bendele R. A. (1993) Electrophoresis 14, 157-161.

17 Andeeson N. L., Esquee-Biasco R., Richaedson F., Foxwoethy P., Eacho P. (1996) Toxicol. Appl. Pharma. 137, 75-89.

18 Dinnen S., Geeich J., Rizza R. (1992) N. Eng. J. Med. 327, 707-713.

19 ADA (1997) Diabetes Care 20, S5-S13.

20 Appel R. D., Sanchez J.-C., Bairoch A., Golaz O., Miu M., Vargas R., Hochstrasser D. (1993) Electrophoresis 14, 12321238.

21 Corbett J. M., Wheeler C. H., Dunn M. J. (1995) Electrophoresis 16, 1524-1529.

22 Liberatori S., Bini L., De Felice C., Magi B., Marzocchi B., Raggiaschi R., Frutiger S., Sanchez J. C., Wilkins M. R., Hughes G., Hochstrasser D. F., Bracci R., Pallini V. (1997) Electrophoresis. 18, 2816-2822.

23 Appel R. D., Bairoch A., Hochstrasser D. F. (1994) Trends Biochem. Sci. 19, 258-260.

24 Appel R. D., Bairoch A., Sanchez J. C., Vargas J. R., Golaz O., Pasquali C., Hochstrasser D. F. (1996) Electrophoresis 17, 540-546.

25 Bairoch A., Apweiler R. (2000) Nucl. Acids Res. 28, 45-48.

26 Baker W., van den Broek A., Camon E., Hingamp P., Sterk P., Stoesser G., Tuli M. A. (2000) Nucl. Acids Res. 28, 19-23.

27 Anderson N. L., Taylor J., Scandora A. E., Coulter B. P., Anderson N. G. (1981) Clin. Chem. 27, 1807-20.

28 Lemkin P. F., Lipkin L. E. (1981) Comput. Biomed. Res. 14, 272-297, 355-380, 407-446.

29 Vo K. P., Miller M. J., Geiduschek E. P., Nielsen C., Olson A., Xuong N. H. (1981) Anal Biochem. 112, 258-71.

30 Garrels J. I. (1989) J. Biol. Chem. 264, 5269-5282.

31 Vincens P., Rabilloud T., Theere H., Pennetier J. L., Ubert M., Tarroux P. (1986) In M. M. Galteau et G. Siest (eds), Recent progress in two-dimensional electrophoresis, Presses Universitäres, Nancy, 121-130.

32 Patton W. F., Lim M. J., Shepro D. (1998) Image acqisition in 2-D electrophoresis. In: 2-D Protocols for Proteome Analysis, A. J. Link (Ed.), Humana Press, Totowa NJ.

33 Armitage P., Berry G.,. (1987) in: Statistical methods and medical research, Blackwell Scientific Publications. Oxford, London.

34 Latter G. I., Burbeck S., Fleming J., Leavitt J. (1984) Clin. Chem. 30, 1925-1932.

35 Eckerskorn A., Jungblut P., Mewes W., Klose J., Lottspeich F. (1988) Electrophoresis 9, 830-838.

36 Hobohm U., Houthaeve T., Sander C. (1994) Anal. Biochem. 222, 202-209.

37 Wilkins M. R., Pasquali C., Appel R. D., Ou K., Golaz O., Sanchez J.-C., Yan J. X., Gooley A. A., Hughes G., Humphery-Smith I., Williams K. L., Hochstrasser D. F. (1996a) Bio/Technology 14, 61-65.

38 Edman P., Begg G. (1967) Eur. J. Biochem. 1, 80-91

39 Henzel W. J., Billeci T. M., Stults J. T., Wong S. C., Grimley C., Watanabe C. (1983) Proc. Natl. Acad. Sci. USA 90, 5011-5015.

40 James P., Quadroni M., Carafoli E., Gonnet G. (1993) Biochem. Biophys. Res. Commun. 195, 58-64.

41 Mann M., Hojrup P., Roepstorff P. (1993) Biol. Mass. Spectrom. 22, 338-345.

42 Pappin D. J. C., Hojeup P., Bleasby A. J. (1993) Curr. Biol. 3, 327-332.

43 Wilkins M. R. and Gooley A. A. (1997c) in Proteome Research: New Frontiers in Functional Genomics, Springer-Verlag Berlin Heidelberg New York, 35-64

44 Bleasby A. J., Akrigg D., Attwood T. K. (1994) Nucl. Acids Res. 22, 3574-3577.

45 Wilkins M. R., Ou K., Appel R. D., Sanchez J.-C., Yan J. X., Goiaz O., Farnsworth V., Cartier P., Hochstrasser D. F., Williams K. L, Gooley A. A. (1996b) Biochem. Biophys. Res. Commun. 221, 609-613.

46 Wilkins M. R., Gasteiger E., Tonelia L., Ou K., Tyler M., Sanchez J.-C., Gooley A. A., Walsh B. J., Bairoch A., Appel R. D., Williams K. L., Hochstrasser D. F. (1998a) J. Mol. Biol. 278, 599-608.

47 Zhang Z., Schaffer A. A., Miller W., Madden T. L., Lipman D. J., Koonin E. V., Altschul S. F. (1998) Nucl. Acids Res. 26, 3986-3990.

48 Gras R., Müller M., Gasteiger E., Gay S., Binz P.-A., Bienvenut W., Hoogiand C., Sanchez J.-C., Bairoch A., Hochstrasser D. F., Appel R. D. (1999) Electrophoresis 20, 3535-3550.

49 Binz P.-A., Wilkins M. R., Gasteiger E., Bairoch A., Appel R. D., Hochstrasser D. F. (1999b) in: Microcharacterization of Proteins, R. Kellner, F. Lottspeich, H. E. Meyer (Eds), Wiley-VCH, 2nd ed., 277-300.

50 Wilkins M. R., Gasteiger E., Wheeler C., Lindskog I., Sanchez J.-C., Bairoch A., Appel R. D., Dunn M. J., Hochstrasser D. F. (1998b) Electrophoresis 19, 3199-3206.

51 Wilkins M. R., Gasteiger E., Gooley A. A., Herbert B. R., Molloy M. P., Binz P.-A., Ou K., Sanchez J.-C., Bairoch A., Williams K. L., Hochstrasser D. F. (1999) J. Mol. Biol. 289, 645-657.

52 Hofmann K., Bucher P., Falquet L., Bairoch A. (1999) Nucl. Acids Res. 27, 215-219.

53 Cooper C. A., Gasteiger E., Packer N. (2001) Proteomics, in press.

54 Packer N. H., Harrison M. J. (1998) Electrophoresis 19, 18721882.

55 Bairoch A. (1997) in Proteome Research: New Frontiers in Functional Genomics, Springer-Verlag Berlin Heidelberg New York, 93-132

56 Benson D. A., Karsch-Mizrachi I., Lipman D. J., Ostell J., Rapp B. A., Wheeler D. L. (2000) Nucl. Acids Res. 28, 15-18

57 Tateno Y., Miyazaki S., Ota M., Sugawara H., Gojobori T. (2000) Nucl. Acids Res. 28, 24-6

58 Eng J. K., McCormack A. L., Yates J. R. 3rd. (1994) J. Am. Soc. Mass Spectrom. 5, 976-989.

59 Adams M. D., Kelley J. M., Gocayne J. D., Dubnick M., Polymeropoulos M. H., Xiao H., Merril C. R., Wu A., Olde B., Moreno R. F., Keriavage A. R., McCombie W. R., Venter J. C. (1991) Science 252, 1651-1656

60 Yates J. R. 3rd, Electrophoresis 1998, 19, 893-900.

61 Bateman A., Birney E., Durbin R., Eddy S. R., Howe K. L., Sonnhammer E. L. L. (2000) Nucleic Acids Res 28, 263-266.

62 Attwood T. K., Croning M. D. R., Flower D. R., Lewis A. P., Mabey J. E., Scordis P., Selley J. N., Wright W. (2000) Nucleic Acids Res. 28, 225-227.

63 Henikoff S., Henikoff J. G., Pietrokovski S. (1999) Bioinformatics 15, 471-479.

64 Corpet 2000 ProDom

65 Apweiler R., Attwood T. K., Bairoch A., Bateman A., Birney E., Biswas M., Bucher P., Cerutti L., Croning M. D. R., Durbin R., Fleischmann W., Hermjakob H., Hulo N., Kahn D., Kanapin A., Karavidopoulou Y., Lopez R., Marx B., Mulder N. J., Oinn T. M., Sigrist C. J. A., Zdobnov E. (2000) submitted to Bioinformatics.

66 Garavelli S. (2000) Nucleic Acids Res. 28, 209-211.

67 R., Birch H., Rapacki K., Brunak S., Hansen J. (1999) Nucleic Acids Res. 27, 370-372.

68 Kreegipuu A., Blom N., Brunak S. (1999) Nucleic Acids Res. 27, 237-239.

69 Molloy M. P., Herbert B. R., Walsh B. J., Tyler M. I., Traini M., Sanchez J.-C., Hochstrasser D. F., Williams K. L., Gooley A. A. (1998) Electrophoresis 19, 837-844.

Target Finding in Genomes and Proteomes

Stefanie Fuhrman, Shoudan Liang, Xiling Wen, and Roland Somogyi

Introduction

The development of novel drug therapies is vital to the advancement of human health. As we move into the twenty-first century, the aging of the population will bring with it a significant rise in degenerative diseases, in particular, Alzheimer's Disease (AD) [1]. Current therapies for AD have only limited effects [2, 3, 4, 5], while the degenerative process eventually results in severe disability. In addition, no method yet exists for the regeneration of tissue.

Recently developed large-scale gene expression assay technologies such as DNA microarrays [6], SAGE (Serial Analysis of Gene Expression) [7], and RT-PCR [8], gene sequence databases such as LifeSeq® database services (Incyte Genomics, Inc.) and GenBank, and gene expression databases such as LifeExpress™ Target (Incyte Genomics, Inc.), now offer the opportunity to discover the logic of tissue degeneration and regeneration. By combining measurements for large numbers of genes expressed in parallel with appropriate data analysis methods, it should be possible to hypothesize functions for newly-discovered genes, redefine functions for familiar genes, and discover the logic of the genetic network.

In this Chapter, we will first discuss why the collection of temporal gene expression data is necessary for the development of therapies for degenerative disorders. We will then describe a straightforward information theoretic method for identifying drug target candidates from among thousands of genes expressed over a time course. Gene expression clustering will be discussed in the context of assigning functions to genes, and a method will be proposed for the combined use of information theory and clustering in therapeutic drug development. Finally, we will discuss the reverse engineering of genetic networks.

Bioinformatics - From Genomes to Drugs. Volume II: Applications. Edited by Thomas Lengauer Copyright © 2002 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-29988-2

120 | 5 Target Finding in Genomes and Proteomes 5.2

Experimental design for large-scale gene expression studies and drug target identification

Current drug therapies for degenerative diseases such as Alzheimer's and Parkinson's Disease are of limited effectiveness, addressing only symptoms while degeneration continues. This should be no surprise given the manner in which degeneration has been approached experimentally, with a focus on individual genes at single stages of disease progression. This approach denies the very nature of degeneration as a process - a series of events. The examination of individual time points provides the equivalent of snapshots of the degenerative process.

Only by collecting data as time series can we observe the whole process and obtain a more complete picture of gene interactions. The technologies and analytical tools needed to accomplish this task are now available. It is now possible to measure gene expression for hundreds or thousands of genes at a time using DNA microarrays [6], SAGE [7], and large-scale RT-PCR [8].

Microarrays, of which there are different types, contain thousands of DNA probes spotted on a surface; each probe is specific for a different gene or EST. Fluorescent-tagged cDNAs from an experimental and a control tissue sample can be hybridized to the probes on the array. The cDNAs of the experimental sample are tagged with a fluorescent probe different from that of the control sample, so that the ratio of experimental to control cDNA can be determined based on color differences.

SAGE [7] involves the concatenation of short nucleotide sequence tags (9-10 bp), each representative of a different gene. The concatenated tags are then cloned and sequenced. The prevalence of a particular tag corresponds to the abundance of the corresponding mRNA in the sample.

RT-PCR can be performed in a semi-quantitative manner [8, 23] for large numbers of genes. This requires a control mRNA, which is reverse-transcribed and amplified along with the sample. Analysis of the resulting PCR products resolved with polyacrylimide gel electrophoresis, involves calculating the ratio of the sample to the control band on the gel. In this way, a relative measure of gene expression can be obtained.

The main benefit of DNA microarrays is their extremely large scale - an entire genome can be represented on a single microarray, while the advantage of SAGE is that only a short sequence is needed to identify each gene. RT-PCR provides the greatest reliability, sensitivity, and dynamic range of all the available large-scale gene expression measurement techniques; its only drawback is the problem of scaling it up with automation.

These techniques are easily employed in generating temporal expression patterns. The only requirement is the proper design of experiments. In that context, appropriate animal models of degenerative diseases must be used,

5.2 Experimental design for large-scale gene expression studies and drug target identification and tissue sampled at different stages of disease progression. The time intervals may be based on the results of published studies that indicate significant changes in gene expression at particular stages of the degenerative process.

Once these multiple-gene, multiple time point data have been collected, it will be necessary to organize them into an intelligible form. This can be accomplished with the right computational tools. Some relatively simple mathematical methods already in use in other fields can be adapted for the analysis of gene expression data. Two such methods discussed here include clustering [8, 9, 10], and Shannon entropy [11, 12].

Figure 5.1 contains an example of a 112-gene temporal expression pattern from the developing rat spinal cord [8], assayed using reverse-transcription polymerase chain reaction (RT-PCR). Genes from a number of recognized functional categories are represented, including a large number of genes involved in intercellular signaling (neurotransmitter receptors, peptides and peptide receptors). The data in Figure 5.1 are displayed so that they may be appreciated by visual inspection. Greater color intensity reflects higher levels of expression. These data have been normalized to maximal expression, so that for each gene, the highest expression detected is 1.0 (maximum color intensity).

The nine time points range from the earliest appearance of the spinal cord at embryonic day 11, to adult (postnatal day 90). Time intervals are based on the rate of development of the rat CNS, which slows as development progresses. It is important to note that time intervals should be spaced appropriately for the system being studied. According to [13], the embryonic rat CNS shows significant anatomical changes at 12 to 24-hour intervals; therefore, to attempt a time series with shorter intervals would be a waste of effort. Similarly, it would be inappropriate to space the time points too far apart, in which case important gene expression events could be missed. In general, published studies demonstrating significant gene expression changes at individual time points can be used as a guide in selecting intervals.

As in the case of normal development, disease progression or recovery from injury may be studied by collecting time series data. This approach is well-suited to the study of animal models of degenerative diseases. It is important to select appropriate animal models - i.e., those that provide a gradual degeneration, rather than the instantaneous destruction of tissue -to mimic human degenerative disease. Similarly, time series of regenerating animal tissue may be studied for clues to inducing this effect in human tissues that do not regenerate.

Wave 1

t— CO LO CO -T— i— 1--1-1— C\lONr-LULU LU LULLI CLCLQ_<

Wave

Wave 1

t— CO LO CO -T— i— 1--1-1— C\lONr-LULU LU LULLI CLCLQ_<

Wave

Clustered temporal patterns for 112 genes expressed in the developing rat spinal cord. Each data point is the average of data from three animals. For each gene, time series data are normalized to maximal expression. The higher the color intensity, the higher the expression value. White means undetectable

Wave 3

i- co in ooi- Tf i—i—i—i—CMON.1— LULU LU LULU0- CL □_<

Clustered temporal patterns for 112 genes expressed in the developing rat spinal cord. Each data point is the average of data from three animals. For each gene, time series data are normalized to maximal expression. The higher the color intensity, the higher the expression value. White means undetectable statin trkC MAP2

synaptophysin neno

GAD65

pre-GAD67

GAT1

ACHE

nAChR 4 QR 2 QR 3 QR 5 GR 1 GR 3 GR 2 mGluR3 mGluR8 NMDA2B

Wave 4

Continue reading here: Constant

Was this article helpful?

0 0