Cancer diagnosis using proteomic patterns

Conrads, Thomas P.; Zhou, Ming; Petricoin, Emanuel F.; Liotta, Lance A.; Veenstra, Timothy D.

doi:10.1586/14737159.3.4.411

Cited by 142 publications

(88 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The approach utilizes classification algorithms to examine spectra in a reproducible manner in order to determine the differences between populations. The most common iteration of this approach is MALDI protein profiling [6][7][8][9][10][11][12][13].…”

Section: Quantitative Relationshipsmentioning

confidence: 99%

Quantitative matrix-assisted laser desorption/ionization mass spectrometry

Duncan¹,

Röder²,

Hunsucker³

2008

Briefings in Functional Genomics and Proteomics

194

162

View full text Add to dashboard Cite

This review summarizes the essential characteristics of matrix-assisted laser desorption/ionization (MALDI) timeof-flight mass spectrometry (TOF MS), especially as they relate to its applications in quantitative analysis. Approaches to quantification by MALDI-TOF MS are presented and published applications are critically reviewed.Keywords: quantification; quantitative analysis; MALDI; mass spectrometry; biomarkers OVERVIEWSince its inception in 1987 [1], matrix-assisted laser desorption/ionization (MALDI) time-of-flight mass spectrometry (TOF MS) has been applied for the analysis of a wide range of biomolecules. Initial applications were almost exclusively for the qualitative analysis of biopolymers because MALDI-TOF MS provided a fast and accurate approach to molecular mass and purity information: Is my material the right mass and is it free from contaminants? Because of the speed of analysis, ease of use, relative low equipment cost, ease of data interpretation and limited potential for cross-contamination between samples/users, MALDI-TOF MS systems were considered walk-up instruments where investigators run their own samples and determine, within minutes, whether they were on the right track with their work.There were, however, two specific areas of application where MALDI-TOF MS was initially viewed as impractical: i.e. the analysis of low mass analytes and quantitative applications. Low mass analysis is complicated by the vast molar excess of the matrix that can swamp any analyte-specific signal in the low m/z region of the spectrum. Quantitative analysis was viewed as implausible because crystallization does not yield a uniform distribution of the analyte and matrix across the target surface and this gives rise to regions where the analyte signal is especially intense relative to other locations on the target surface. MALDI MS was therefore viewed as inherently irreproducible because, for a given amount of analyte loaded onto the target, the measured ion intensity varies.Subsequently, it has been shown that both these perceived limitations can be overcome and quantification can be routine, both for low and high mass analytes. Now, an extensive list of published applications includes quantification of amino acids, lipids, natural products, drugs, polymers, herbicides, metabolites, toxins, oligonucleotides, carbohydrates, peptides and proteins. (Selected applications from these areas are discussed at the end of this review.) Here, we focus on the quantitative applications of MALDI and illustrate applications relating to both low and high-molecular mass analytes. We also discuss the strategies for applying MALDI to relative and absolute quantification. While MALDI is most commonly combined with TOF analysers, other analyser options are possible and applications employing these are also presented. Our aim is not to provide an exhaustive catalogue of published applications, but to discuss the general principles and to illustrate the

show abstract

Section: Quantitative Relationshipsmentioning

confidence: 99%

Quantitative matrix-assisted laser desorption/ionization mass spectrometry

Duncan¹,

Röder²,

Hunsucker³

2008

Briefings in Functional Genomics and Proteomics

194

162

View full text Add to dashboard Cite

show abstract

“…This further supported the claim that down-sampling appears detrimental to classification accuracy. The conclusions drawn are in line with those in (Conrads et al, 2003) where changes in resolution created by different MS techniques produced similar results. Because the MS spectra are histograms describing the ion concentrations based on the mass-to-charge ratios, the low resolution techniques effectively aggregate distinct ion concentrations into a Classification accuracy on progressively down-sampled data.…”

Section: Dimensionality Reductionsupporting

confidence: 88%

“…Although cross-validation studies were not conducted, the approach was able to correctly classify all cancer stricken patients and 95% of healthy women, on a single test set. Motivated by the need for greater recall and precision, in (Conrads et al, 2003), a low resolution mass spectrometry technique was compared with a high resolution technique using the same ovarian cancer data set. The goal was to determine whether sensitivity and PPV ‡ (i.e., recall and precision) scores would improve by using a higher resolution spectra provided by the SELDI TOF MS hardware § .…”

Section: Ovarian Cancer Studiesmentioning

confidence: 99%

“…This technique, essentially removes high frequency components. In order to test the conjecture made in (Conrads et al, 2003), that higher resolution data tends to improve classification performance, we will use this approach to test the merit of dimensionality reduction via down sampling.…”

Section: Dimensionality Reductionmentioning

confidence: 99%

“…In addition, for many diseases it is suspected that no single biomarkers exit, which are capable of producing reliable diagnoses. (Conrads et al, 2003) While high-resolution mass spectrometry techniques are thought to have potential for accurate diagnosis due to the vast amount of information captured, they are problematic for supervised training of classifiers. Specifically, the many thousands of raw attributes forming the spectra frequently contain a large amount of redundancy, information irrelevant to a particular disease, and measurement noise.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Feature Extraction for Classification of Proteomic Mass Spectra: A Comparative Study

View full text Add to dashboard Cite

Summary.To satisfy the ever growing need for effective screening and diagnostic tests, medical practitioners have turned their attention to high resolution, high throughput methods. One approach is to use mass spectrometry based methods for disease diagnosis. Effective diagnosis is achieved by classifying the mass spectra as belonging to healthy or diseased individuals. Unfortunately, the high resolution mass spectrometry data contains a large degree of noisy, redundant and irrelevant information, making accurate classification difficult. To overcome these obstacles, feature extraction methods are used to select or create small sets of relevant features. This paper compares existing feature selection methods to a novel wrapper-based feature selection and centroid-based classification method. A key contribution is the exposition of different feature extraction techniques, which encompass dimensionality reduction and feature selection methods. The experiments, on two cancer data sets, indicate that feature selection algorithms tend to both reduce data dimensionality and increase classification accuracy, while the dimensionality reduction techniques sacrifice performance as a result of lowering the number of features. In order to evaluate the dimensionality reduction and feature selection techniques, we use a simple classifier, thereby making the approach tractable. In relation to previous research, the proposed algorithm is very competitive in terms of (i) classification accuracy, (ii) size of feature sets, (iii) usage of computational resources during both training and classification phases.

show abstract