生物谷報(bào)道：分子掃描是一種全新的生物技術(shù)方法,，是最近才發(fā)展起來(lái)的。請(qǐng)看相關(guān)的全文報(bào)道：

The molecular scanner: concept and developments
Current Opinion in Biotechnology, 2004, 15:1:17-23

Pierre-Alain Binz , 1, 2, Markus Müller 1, Christine Hoogland 1, Catherine Zimmermann 3, Carla Pasquarello 3, Garry Corthals 3, Jean-Charles Sanchez 2, 3, Denis F Hochstrasser 2, 3 and Ron D Appel 1, 2, 3

1 Swiss Institute of Bioinformatics, Proteome Informatics Group, CMU, Michel Servet 1, 1211, Geneva, Switzerland
2 Geneva University Faculty of Medicine, 1211, Geneva, Switzerland
3 Biomedical Proteomics Research Center, Geneva University Hospital, 1211, Geneva, Switzerland

Available online 9 January 2004.

Introduction
Proteomics is the analysis of proteomes (i.e. the protein complement to the genome) and necessitates the study of complete sets of proteins. Currently, no single technology has been shown to be comprehensive in the analysis of all proteins of a biological sample. There are many reasons for this difficulty, and to understand the issues to be solved in proteomics one needs to first understand the complexity of the samples under study [1.]. Variations in protein concentrations, the number of different protein forms and physicochemical diversity all have to be taken into consideration.

The challenges of sample complexity
Concentration
If one considers the dynamic range of protein concentrations found in human plasma there is a large variation, from the millimolar range (e.g. for albumin) down to femtomolar concentrations (e.g. for tumour necrosis factor), which represents a difference of 12 logs. If we also consider the case where cells die and release their contents into the bloodstream, it might be expected that a couple of copies of a rare transcription factor could exist in four litres of plasma, extending the concentration range to 21 logs [2.].

The number of different protein forms
Many genes have been shown to produce multiple mRNAs and proteins through several mechanisms, including gene splicing, mRNA editing and modifications at the co-translational and post-translational levels. One can anticipate that, on average, in humans one gene produces five or six final protein products. If the human genome contains 35,000 genes, one might expect more than 200,000 different proteins in the human body. Even if all of these proteins are not found in all tissues, cells, organelles or biological fluids, many of them might be found in the blood, but at extremely low concentrations. Our previous work on the plasma proteome at the University Hospital in Geneva and at Geneprot (http://www.geneprot.com) confirms this hypothesis [2.].

Physicochemical diversity
Proteins are designed to act in various media. They can be highly soluble to reach millimolar concentrations in the blood, but in other cases they have to be stable in highly hydrophobic environments, such as cellular membranes. The total size of characterized proteins ranges from a couple of dozen amino acids (about 2500 proteins of 50 amino acids or less are annotated in Swiss-Prot knowledgebase [3.]) to more than 10,000 amino acids (e.g. human Nesprin 1, a nuclear envelope protein, is annotated with potentially 8746 amino acids in Swiss-Prot [accession number Q8NF91] and human titin 1 is estimated to have a length of 34,350 amino acids in TrEMBL [3.] [accession number Q8WZ42]). Besides their individual differences, active proteins also interact with other proteins in the form of molecular complexes. Likewise, they interact with DNA or RNA when they act as activators and with small molecules when their function is enzymatic. The structure of such complexes modifies the physicochemical behaviour of their components.

Functional challenges in proteomics
Against the background of this complexity, many experimental methods have been developed that try to overcome the current challenges in proteomics. The functional challenges of proteomics that need to be addressed fall into four categories, discussed below.

Firstly, it is necessary to identify the proteins contained in a sample. In most cases, this is understood as matching experimental parameters extracted from isolated and processed proteins or peptides to sequence database entries, starting from a given biological material. Secondly, the different protein forms contained in a sample must be identified. This completes the protein identification by decrypting the detailed structure of the polypeptides present in a sample. This step includes searching for products of splicing or truncation events and identifying the presence of post-translational modifications, and so on. The third challenge is to process and visualize proteomes using different methods that can represent one or more proteome(s) as separated proteins or peptides in a two- (or more) dimensional manner. Finally, proteomics needs to identify differences between two or more samples. This not only implies a search for the presence or absence of some polypeptides in one sample compared with another, but also includes quantitative measurements of the variation of protein expression levels between two or more samples (or between two or more populations of samples).

There is a definite need to establish methods that can handle such complex material in a robust manner, with reasonably high throughput, high sensitivity and specificity. In addition to the currently used and well-described experimental workflows that include classical gel-digestion mass spectrometry (MS), the multidimensional protein identification technology (MuDPIT) method [4.], the isotope-coded affinity tag (ICAT) quantitation approach [5.], surface-enhanced laser desorption ionisation (SELDI) technology [6.], and some protein and peptide array techniques [7. and 8.] and their derivatives, the molecular scanner addresses most of the above described challenges in one innovative manner.

In this review we describe the concept behind the molecular scanner and discuss approaches to data handling. We also consider the advantages and disadvantages of this methodology for proteomic studies and look at future developments.

The concept of the molecular scanner
In principle, the concept of the molecular scanner was first described and predicted in 1991 as a method to visualize and classify biological samples at a molecular discriminating level, using the power of two-dimensional SDS polyacrylamide gel electrophoresis (PAGE) [9.]. With the development of bioinformatics and mass spectrometry, the molecular scanner has been refined and today aims to maximize the benefit of parallel protein transfer during western blotting and the imaging power and potential of a mass spectrometer.

The original ideas behind the concept of the molecular scanner are summarized in Box 1. The ultimate aim is to develop a highly automated and highly specific approach to generating a multidimensional representation of an annotated proteome.

Box 1. The concept behind the molecular scanner

Experimental implementation

In the original description of the molecular scanner (Figure 1) [10. and 11.], an enzymatic membrane was developed where, for example, trypsin was covalently attached to a hydrophilic membrane. The enzymatic membrane is then intercalated between a gel on which proteins have been separated and a collecting membrane. The attached enzyme processes most proteins from the gel during the electrotransfer through the membrane onto the collecting surface. The final support that contains the modified (digested) peptides is covered with an appropriate matrix and scanned by the UV or IR laser of a matrix-assisted laser desorption ionization (MALDI) mass spectrometer. Software tools reconstruct the image of the original gel or tissue with multidimensional views and provide for the automatic identification and partial characterization of the proteins separated in the gel. Data correlation algorithms take advantage of the spatial resolution and distribution of the MS signal, enhancing significantly the signal-to-noise ratio. Scanning bacterial and human samples that were separated by one-dimensional and two-dimensional gel electrophoresis revealed the sensitivity and power of this technique [11., 12. and 13.].
 
 (38K)

Figure 1. Conceptual description of the molecular scanner workflow.

The application of this approach has led to several dedicated developments and optimizations. These include the development of new enzyme-bound membranes, the optimization of experimental conditions, and the development of new algorithms and protein identification tools. Box 2 summarises many of the recent improvements.

Box 2. The development and optimization of the molecular scanner

The experimental implementation is flexible and open to variation

The starting material can be either a gel produced by one- (1-DE) or two-dimensional electrophoresis (2-DE) or a fine slice of tissue. Attempts to process peptides and proteins directly from a frozen tissue section, electrotransferred through a trypsin-membrane, have already shown encouraging results (M Stoeckli, personal communication; C Zimmermann, personal communication).
The technology is able to accept binary mixtures. Isotopically labelled samples can be mixed before the separation technique is chosen (1-DE or 2-DE). This allows samples to be compared and, thus, enables the relative quantitation of biological material.
The enzyme membrane was originally developed with porcine trypsin. Some attempts have been made to replace trypsin with other enzymes, and, in principal, one can think of many endoproteases that could be used (such as Lys-C, protease V8, etc.) or other functional enzymes, such as alkaline phosphatase. It should be noted, however, that the choice of enzyme can have an impact on the conditions used for electrotransfer; the enzyme should be compatible with the transfer buffer.
The experimental conditions for electrotransfer have been worked out for various approaches. The most direct method is the one-step digestion transfer (OSDT), where the entire protein sample is extracted from gels and submitted to the enzymatic membrane. The optimal conditions currently in use are based on a square-shaped tension (i.e. an alternating tension of the type 125 ms at +12V then 125 ms at -5V). The double-parallel digestion approach combines the OSDT with a `pre-digestion' step, where a minimal amount of trypsin is introduced into the original gel. This additional step allows for some proteolytic modification (i.e. predigest) of those proteins that are difficult to transfer as entire molecules (e.g. very large or very basic proteins).
The MALDI matrix has to be deposited on the collecting membrane in such a way that peptides can be extracted from the membrane surface but do not have to diffuse from their original locations. Currently, a mechanical spray is the most convenient way to apply matrix solution to a collecting membrane. For the analysis of small molecules, the standard use of the matrix solution HCCA (-cyano-cinnamic acid) can be replaced with some other matrix molecules, such as sinapinic acid or DHBA (2,5-dihydroxybenzoic acid). The membrane itself can be fixed on a modified MALDI plate either with high vacuum grease or with double-sided carbon tape.
The original acquisition mode was a systematic and standard peptide mass fingerprinting (PMF) measurement at predefined positions on the membrane. Currently available MALDI-TOF-TOF (MALDI-time-of-flight-time-of-flight) or MALDI-Q-TOF (MALDI-quadrupole-time-of-flight) MS instruments have shown the possibility to acquire MS/MS spectra from a collecting membrane [14.]. Therefore, it is possible to acquire a first rapid MS scan of the whole surface, which can then be analysed. The positions of detected signals can then be chosen to measure precise PMF spectra. After clustering of masses and/or identification of proteins, a third scan can be performed. This tandem mass spectrometry (MS/MS) scan would focus on the acquisition of tandem mass spectra for particularly interesting precursor ions. These can represent either MS signals to validate or MS signals for which no identification was possible.
Bioinformatics treatment of the generated data

The data processing takes advantage of the redundancy and neighbourhood of MS signals to filter noise and to select confident signals (Figure 2).
 
 (22K)

Figure 2. Dataflow in the molecular scanner software.

In the analysis of proteins separated by 2-DE, visual inspection of the positional distribution of the total ion count does not reveal some weakly expressed spots, owing to the presence of intense large spots and rather intense chemical noise. Much more information can be found, however, by looking at the selected ion counts (SIC) (i.e. the two-dimensional intensity distribution of a specific mass). Visualization of all peptide mass fingerprint data shows that some masses are localized in spots, whereas others, especially in the lower mass region, spread out over the entire membrane. These masses can be attributed to chemical noise (e.g. matrix clusters, trypsin peptides and human keratin) and can be discarded from the mass fingerprints. A method that clusters peptide masses according to the similarity of the spatial distributions of their SIC has been presented. This method allowed spot detection that was much more sensitive than Coomassie blue staining [11.]. The approach could also be used to improve PMF identification scoring, as a protein present in a spot should also have peptide masses with an SIC localized in the same spot. This method allowed many of the false positives that usually go along with PMF identifications to be discarded and enabled many weakly expressed proteins present in the gel to be identified. In an experiment with human plasma most of the proteins that are annotated in the SWISS2D-PAGE [15. and 16.] database could be clearly identified, despite the low resolution of the scan (~1200 Da in MW direction) and the presence of albumin and other abundant proteins [12.]. These results would not have been possible without making use of the data-processing methods described above.

If data acquisition is extended to MS/MS with a MALDI-TOF-TOF instrument, a first scan in the MS mode could detect spots and non-matching masses, which could then be further analysed in the MS/MS mode [2.].

The identification process was originally made possible through the optimization of an algorithm of the PMF tool SmartIdent [13.]. Today, optimized tools such as Aldente [17.] (for PMF) or OLAV [18.] (for MS/MS) look promising for providing the high-throughput required by the rapid acquisition rate of the current MS instrumentation.

Data visualization
High-throughput methods can produce a large amount of MS data and the multidimensional visualization of this data is becoming increasingly important (Figure 3). Such visualization allows the data to be surveyed and provides ideas for algorithmic solutions. One example is provided by secondary ion mass spectrometry (SIMS) techniques, where natural tissues can be scanned with a spatial resolution of less than 100 nm and the resulting spectra can be used to visualize the two- or three-dimensional distributions of secondary ions [19.]. Stoeckli et al. [20. and 21.] coated frozen thin sections of tissue with a solution of MALDI matrix, then dried and introduced them into a mass spectrometer that scanned the sample. The position of user-defined mass values (m/z values) could be visualized and matched with subcellular positioning.

 
 (67K)

Figure 3. The identification of Escherichia coli proteins. Various methods can be used to visualise molecular scanner data. The area represented here corresponds to 1536 MS spectra measured on a surface of about 1.2 cm2 from a mini-two-dimensional electrophoretic gel of E. coli. (a) A total MS intensity image. The spots were detected with the Melanie image analysis software. (b) The smoothed spot centres calculated with the molecular scanner software. (c) A positioning of the protein identifications on the E. coli surface analysed. (d,e) The positions and MS intensities of the m/z values 951,627 Da and 1108,664 Da on the whole surface, respectively. (f) Examples of the localisation of some of the identified proteins. The four columns contain the Swiss-Prot ID, the Swiss-Prot AC, the location of the identified protein (with intensities represented as the sum of the intensities of the mass cluster elements), and the list of the m/z values in the corresponding mass clusters. The underlined masses match a peptide mass of the identified protein. The masses marked with an asterisk match a potential post-translational modification in FindMod (http://www.expasy.org/tools/). 6-Phosphogluconate dehydrogenase (6PGD, marked with an asterisk) is an example of a protein that is not identified in SWISS-2DPAGE. IDH, isocitrate dehydrogenase; PGK, phosphoglycerate kinase; METK, S-adenosylmethionine synthetase; the dashed lines link the representations of METK in the various reconstructed images.

As an extension of this approach, the molecular scanner not only generates m/z values as a function of their position on a scanned surface [22.], but also generates positional discrimination of mass clusters. These can be represented as reconstituted spots or bands with intensities corresponding to the intensities of the represented m/z values in the spectra. As the identification of proteins and peptides is obtained with associated m/z values, their positional description can be visualized with reconstituted spots the intensities of which are a function of the identification scores, number of peptides identified, and so on.

Interconnection with databases
One single molecular scanner experiment can generate a lot of data. If these data are stored in a database, various questions can be addressed. For instance, what is the list of identified proteins? How many mass clusters represent the protein ABC? Does this spot contain more than one protein? What are the m/z signals, specific for a protein, that have not been identified?

The automatic analysis of molecular data generates annotated images that could be included in a public database, such as SWISS-2D-PAGE [13. and 14.].

Advantages and limitations of the technology
As with all currently available methodologies and technologies in proteomics, the molecular scanner has advantages and disadvantages over other existing approaches.

Advantages
One of they key advantages of the molecular scanner is that minimal sample handling is required. A single protein sample is used as a starting point (e.g. one gel or one slice of tissue) and this material is processed as a whole, with the sample being transferred to the collecting membrane in a single step. There is no need to excise spots or to chemically stain the material.

Once on the collecting membrane, the sample is stable and available for analysis for a couple of days. This means that, at the end of an MS analysis, all peptides are still available for further MS or MS/MS data acquisition. Furthermore, owing to the small size of the MALDI laser beam (about 50–100 m), a protein spot can be measured more than once, at different locations. Therefore, the method is able to discriminate spectra at various regions of a spot, and, as a consequence, deconvolute overlapping spots.

Limitations
The comprehensivity of the molecular scanner is strongly dependent on the wet-laboratory technologies used for protein separation. In cases where 2-DE is used, the method will obviously be unable to identify proteins that do not migrate in this medium; 1-DE might efficiently replace 2-DE for 2-DE-difficult proteins. When the analysis is based on PMF spectra, the peptide density at each data point might be too high and generate spectra that are too crowded when a complex sample is separated by 1-DE. The limitation here is inherent to the limitation of PMF to identify more than a couple of proteins at each data point. Employing protein prefractionation or protein chromatographic separation followed by one-dimensional SDS–PAGE can improve the resolution.

The sensitivity limit of the method is quantified in terms of the amount of peptide and/or protein that can be detected per surface unit. The method does not allow an entire protein spot or band to be concentrated in a small volume; however, using this approach proteins have already been identified that are not visible by gel staining.

Conclusions
The development of the molecular scanner has revealed the extremely important need for tight interactions between wet-laboratory scientists, bioinformaticians, clinicians and industrial hardware providers to bring innovative proteomics technologies to fruition. As an example of a fruitful collaboration, technology transfer and commercialization of part of the molecular scanner concept, an experimental kit is under development at Applied Biosystems (http://www.appliedbiosystems.com/).

In addition to its role as a method for parallel protein identification and molecular imaging, the molecular scanner can be seen as an experimental and bioinformatics environment that approaches the analysis of a proteome or proteomes using multidimensional representations. These representations can be chosen for specific end-users.

The molecular scanner in its current form has shown the power of this methodology in terms of the ease of sample handling, sensitivity, specificity, and capacity to perform efficient and informative molecular imaging. It also offers possibilities for future developments, including applications in the clinical environment. For example, clinical applications are currently being developed to diagnose proteinuria and solid tumours from urine and biopsy specimens, respectively.

New chemical processing methods and mass spectrometers with improved sensitivity and throughput are likely to extend and enhance the use of the molecular scanner in the years to come.

References and recommended reading
Papers of particular interest, published within the annual period of review, have been highlighted as:

 of special interest
 of outstanding interest

 
References
1.. D.F. Hochstrasser, J.C. Sanchez and R.D. Appel, Proteomics and its trends facing nature's complexity. Proteomics 2 (2002), pp. 807–812. Abstract-Elsevier BIOBASE | Abstract-EMBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

A review that discuss the challenges for proteomics to approach the analysis of the complexity inherent to biological samples.

2.. Hochstrasser DF, Corthals G, Zimmermann C, Scherl A, Bougueleret L, Colinge J, Masselot A, Binz P-A, Muller M, Bairoch A et al.: Proteomics & Mass Spectrometry in Medicine. In Proceedings of the 2003 International Mass Spectrometry Society conference. Edinburgh; UK: in press.

A review of the sentitive and challenging issues of current proteomics applied to medicine, with a focus on the use of mass spectrometry.

3. B. Boeckmann, A. Bairoch, R. Apweiler, M.-C. Blatter, A. Estreicher, E. Gasteiger, M.J. Martin, K. Michoud, C. O'Donovan, I. Phan et al., The Swiss-Prot protein knowledgebase and its supplement TrEMBL in 2003. Nucleic Acids Res. 31 (2003), pp. 365–370. Abstract-EMBASE | Abstract-Elsevier BIOBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

4. D.A. Wolters, M.P. Washburn and J.R. Yates, III, An automated multidimensional protein identification technology for shotgun proteomics. Anal. Chem. 73 (2001), pp. 5683–5690. Abstract-EMBASE | Abstract-MEDLINE | Abstract-Beilstein Abstracts   | $Order Document | Full Text via CrossRef

5. S.P. Gygi, B. Rist, S.A. Gerber, F. Turecek, M.H. Gelb and R. Aebersold, Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nat. Biotechnol. 17 (1999), p. 994. Abstract-Elsevier BIOBASE | Abstract-EMBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

6. M. Merchant and S.R. Weinberger, Recent advancements in surface-enhanced laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 21 (2000), pp. 1164–1177. Abstract-EMBASE | Abstract-Elsevier BIOBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

7. D.J. Cahill and E. Nordhoff, Protein arrays and their role in proteomics. Adv. Biochem. Eng. Biotechnol. 83 (2003), pp. 177–187. Abstract-MEDLINE   | $Order Document

8. R.E. Jenkins and S.R. Pennington, Arrays for protein expression profiling: towards a viable alternative to two-dimensional gel electrophoresis?. Proteomics 1 (2001), pp. 13–29. Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

9. D.F. Hochstrasser, R.D. Appel, R. Vargas, R. Perrier, J.F. Vurlod, F. Ravier, C. Pasquali, M. Funk, C. Pellegrini, A.F. Muller and M.D. Scherrer, Clinical molecular scanner: the Melanie project. MD Comput. 8 (1991), p. 85. Abstract-MEDLINE | Abstract-INSPEC | Abstract-EMBASE   | $Order Document

10. W.V. Bienvenut, J.-C. Sanchez, A. Karmime, V. Rouge, K. Rose, P.-A. Binz and D.F. Hochstrasser, Toward a clinical molecular scanner for proteome research: parallel protein chemical processing before and during westernblot. Anal. Chem. 71 (1999), pp. 4800–4807. Abstract-Beilstein Abstracts | Abstract-EMBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

11. P.-A. Binz, M. Müller, D. Walther, W.V. Bienvenut, R. Gras, C. Hoogland, G. Bouchet, E. Gasteiger, R. Fabbretti, S. Gay et al., A molecular scanner to highly automate proteomic research and to display proteome images. Anal. Chem. 71 (1999), pp. 4981–4988. Abstract-MEDLINE | Abstract-Beilstein Abstracts | Abstract-EMBASE   | $Order Document | Full Text via CrossRef

12.. M. Muller, R. Gras, P.-A. Binz, D.F. Hochstrasser and R.D. Appel, Improving protein identification for a molecular scanner experiment with human plasma by using correlations between spectra. Proteomics 10 (2002), pp. 1413–1425. Abstract-MEDLINE   | $Order Document

Consideration of correlation of similarities and differences between PMF spectra from neighbouring positions in a molecular scanner experiment is improving quality and specificity of protein identification in complex mixtures.

13. R. Gras, M. Muller, E. Gasteiger, S. Gay, P.A. Binz, W. Bienvenut, C. Hoogland, J.C. Sanchez, A. Bairoch, D.F. Hochstrasser and R.D. Appel, Improving protein identification from peptide mass fingerprinting through a parameterized multi-level scoring algorithm and an optimized peak detection. Electrophoresis 20 (1999), pp. 3535–3550. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Elsevier BIOBASE   | $Order Document | Full Text via CrossRef

14. Nadler T, Huang Y, Wagenfeld B, Parker B, Lotti R, Vella GJ, Binz P-A, Müller M, Gras R, Hochstrasser DF et al.: Proteome Analysis by Multi-Dimensional Liquid Chromatography Gel-Electrophoresis and the Molecular Scanner. Presented at the 5th Siena 2D electrophoresis meeting: Siena, Italy; September 1–5, 2002.

15. C. Hoogland, J.C. Sanchez, L. Tonella, P.A. Binz, A. Bairoch, D.F. Hochstrasser and R.D. Appel, The 1999 SWISS-2DPAGE database update. Nucleic Acids Res. 28 (2000), pp. 286–288. Abstract-EMBASE | Abstract-Elsevier BIOBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

16. E. Gasteiger, A. Gattiker, C. Hoogland, I. Ivanyi, R.D. Appel and A. Bairoch, ExPASy: The proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 31 (2003), pp. 3784–3788. Abstract-Elsevier BIOBASE   | $Order Document | Full Text via CrossRef

17. Tuloup M, Hernandez C, Coro I, Hoogland C, Binz P-A, Appel RD: Aldente and BioGraph: an improved peptide mass fingerprinting protein identification environment. In Proceedings of the Swiss Proteomics Society 2003 Congress, Understanding biological systems through proteomics. Basle, 2–4 December 2003; Fontis Media 2003, pp 174–176.

18. J. Colinge, A. Masselot, M. Giron, T. Dessingy and J. Magnin, OLAV: towards high-throughput tandem mass spectrometry data identification. Proteomics 3 (2003), pp. 1454–1463. Abstract-EMBASE | Abstract-Elsevier BIOBASE   | $Order Document | Full Text via CrossRef

19. M.L. Pacholski and N. Winograd, Imaging with mass spectrometry. Chem. Rev. 99 (1999), pp. 2977–3005.

20. M. Stoeckli, P. Chaurand, D.E. Hallahan and R.M. Caprioli, Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nat. Med. 7 (2001), pp. 493–496. Abstract-Elsevier BIOBASE | Abstract-EMBASE | Abstract-MEDLINE   | $Order Document | Full Text via CrossRef

21.. M. Stoeckli, D. Staab, M. Staufenbiel, K.H. Wiederhold and L. Signor, Molecular imaging of amyloid peptides in mouse brain sections using mass spectrometry. Anal. Biochem. 311 (2002), pp. 33–39. SummaryPlus | Full Text + Links | PDF (424 K)

An MS-based approach to analyse the subcellular location of polypeptides from tissue sections.

22.. M. Muller, R. Gras, R.D. Appel, W.V. Bienvenut and D.F. Hochstrasser, Visualization and analysis of molecular scanner peptide mass spectra. J. Am. Soc. Mass Spectrom. 13 (2002), pp. 221–231. SummaryPlus | Full Text + Links | PDF (412 K)

This paper decribes the data analysis issues and the visual representation possibilities of