Machine learning approaches in MALDI-MSI: clinical applications

Introduction: Despite the unquestionable advantages of Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Imaging in visualizing the spatial distribution and the relative abundance of biomolecules directly on-tissue, the yielded data is complex and high dimensional. Therefore, analysis and interpretation of this huge amount of information is mathematically, statistically and computationally challenging.

Areas covered: This article reviews some of the challenges in data elaboration with particular emphasis on machine learning techniques employed in clinical applications, and can be useful in general as an entry point for those who want to study the computational aspects. Several characteristics of data processing are described, enlightening advantages and disadvantages. Different approaches for data elaboration focused on clinical applications are also provided. Practical tutorial based upon Orange Canvas and Weka software is included, helping familiarization with the data processing.

Expert commentary: Recently, MALDI-MSI has gained considerable attention and has been employed for research and diagnostic purposes, with successful results. Data dimensionality constitutes an important issue and statistical methods for information-preserving data reduction represent one of the most challenging aspects. The most common data reduction methods are characterized by collecting independent observations into a single table. However, the incorporation of relational information can improve the discriminatory capability of the data.     

Integration of mRNA Expression Profile,Copy Number Alterations,and microRNA Expression Levels in Breast Cancer to Improve Grade Definition

Defining the aggressiveness and growth rate of a malignant cell population is a key step in the clinical approach to treating tumor disease. The correct grading of breast cancer (BC) is a fundamental part in determining the appropriate treatment. Biological variables can make it difficult to elucidate the mechanisms underlying BC development. To identify potential markers that can be used for BC classification, we analyzed mRNAs expression profiles, gene copy numbers, microRNAs expression and their association with tumor grade in BC microarray-derived datasets. From mRNA expression results, we found that grade 2 BC is most likely a mixture of grade 1 and grade 3 that have been misclassified, being described by the gene signature of either grade 1 or grade 3. We assessed the potential of the new approach of integrating mRNA expression profile, copy number alterations, and microRNA expression levels to select a limited number of genomic BC biomarkers. The combination of mRNA profile analysis and copy number data with microRNA expression levels led to the identification of two gene signatures of 42 and 4 altered genes (FOXM1, KPNA4, H2AFV and DDX19A) respectively, the latter obtained through a meta-analytical procedure. The 42-based gene signature identifies 4 classes of up- or down-regulated microRNAs (17 microRNAs) and of their 17 target mRNA, and the 4-based genes signature identified 4 microRNAs (Hsa-miR-320d, Hsa-miR-139-5p, Hsa-miR-567 and Hsa-let-7c). These results are discussed from a biological point of view with respect to pathological features of BC. Our identified mRNAs and microRNAs were validated as prognostic factors of BC disease progression, and could potentially facilitate the implementation of assays for laboratory validation, due to their reduced number.     

Different expression of Fibrinopeptide A and related fragments in serum of type 1 diabetic patients with nephropathy

Type 1 diabetes (insulin-dependent diabetes mellitus, IDDM) is an autoimmune disease affecting about 0.12% of the world's population. Diabetic nephropathy (DN) is a major long-term complication of both types of diabetes and retains a high human, social and economic cost. Thus, the identification of markers for the early detection of DN represents a relevant target of diabetic research.The present work is a pilot study focused on proteomic analysis of serum of controls (n = 9), IDDM patients (n = 10) and DN patients (n = 4) by the ClinProt profiling technology based on mass spectrometry. This approach allowed to identify a pattern of peptides able to differentiate the studied populations with sensitivity and specificity close to 100%. Variance of the results allowed to estimate the sample size needed to keep the expected False Discovery Rate low. Moreover, three peptides differentially expressed in the serum of patients as compared to controls were identified by LC-ESI MS/MS as the whole fibrinopeptide A peptide and two of its fragments, respectively. The two fragments were under-expressed in diabetic patients, while Fibrinopeptide A was over-expressed, suggesting that anomalous turnover of Fibrinopeptide A could be involved in the pathogenesis of DN.     

Combined analysis of chromosomal instabilities and gene expression for colon cancer progression inference

Background

The human body plays host to a vast array of bacteria, found in oral cavities, skin, gastrointestinal tract and the vagina. Some bacteria are harmful while others are beneficial to the host. Despite the availability of many methods to identify bacteria, most of them are only applicable to specific and cultivable bacteria and are also tedious. Based on high throughput sequencing technology, this work derives 16S rRNA sequences of bacteria and analyzes probiotics and pathogens species.

Results

We constructed a database that recorded the species of probiotics and pathogens from literature, along with a modified Smith-Waterman algorithm for assigning the taxonomy of the sequenced 16S rRNA sequences. We also constructed a bacteria disease risk model for seven diseases based on 98 samples. Applicability of the proposed platform is demonstrated by collecting the microbiome in human gut of 13 samples.

Conclusions

The proposed platform provides a relatively easy means of identifying a certain amount of bacteria and their species (including uncultivable pathogens) for clinical microbiology applications. That is, detecting how probiotics and pathogens inhabit humans and how affect their health can significantly contribute to develop a diagnosis and treatment method.     

Urinary Signatures of Renal Cell Carcinoma Investigated by Peptidomic Approaches

Renal Cell Carcinoma (RCC) is typically asymptomatic and surgery usually increases patient''s lifespan only for early stage tumours. Moreover, solid renal masses cannot be confidently differentiated from RCC. Therefore, markers to distinguish malignant kidney tumours and for their detection are needed. Two different peptide signatures were obtained by a MALDI-TOF profiling approach based on urine pre-purification by C8 magnetic beads. One cluster of 12 signals could differentiate malignant tumours (n = 137) from benign renal masses and controls (n = 153) with sensitivity of 76% and specificity of 87% in the validation set. A second cluster of 12 signals distinguished clear cell RCC (n = 118) from controls (n = 137) with sensitivity and specificity values of 84% and 91%, respectively. Most of the peptide signals used in the two models were observed at higher abundance in patient urines and could be identified as fragments of proteins involved in tumour pathogenesis and progression. Among them: the Meprin 1α with a pro-angiogenic activity, the Probable G-protein coupled receptor 162, belonging to the GPCRs family and known to be associated with several key functions in cancer, the Osteopontin that strongly correlates to tumour stages and invasiveness, the Phosphorylase b kinase regulatory subunit alpha and the SeCreted and TransMembrane protein 1.     

From protein-protein interactions to protein co-expression networks: a new perspective to evaluate large-scale proteomic data

The reductionist approach of dissecting biological systems into their constituents has been successful in the first stage of the molecular biology to elucidate the chemical basis of several biological processes. This knowledge helped biologists to understand the complexity of the biological systems evidencing that most biological functions do not arise from individual molecules; thus, realizing that the emergent properties of the biological systems cannot be explained or be predicted by investigating individual molecules without taking into consideration their relations. Thanks to the improvement of the current -omics technologies and the increasing understanding of the molecular relationships, even more studies are evaluating the biological systems through approaches based on graph theory. Genomic and proteomic data are often combined with protein-protein interaction (PPI) networks whose structure is routinely analyzed by algorithms and tools to characterize hubs/bottlenecks and topological, functional, and disease modules. On the other hand, co-expression networks represent a complementary procedure that give the opportunity to evaluate at system level including organisms that lack information on PPIs. Based on these premises, we introduce the reader to the PPI and to the co-expression networks, including aspects of reconstruction and analysis. In particular, the new idea to evaluate large-scale proteomic data by means of co-expression networks will be discussed presenting some examples of application. Their use to infer biological knowledge will be shown, and a special attention will be devoted to the topological and module analysis.     

Availability of MudPIT data for classification of biological samples

Background

Mass spectrometry is an important analytical tool for clinical proteomics. Primarily employed for biomarker discovery, it is increasingly used for developing methods which may help to provide unambiguous diagnosis of biological samples. In this context, we investigated the classification of phenotypes by applying support vector machine (SVM) on experimental data obtained by MudPIT approach. In particular, we compared the performance capabilities of SVM by using two independent collection of complex samples and different data-types, such as mass spectra (m/z), peptides and proteins.

Results

Globally, protein and peptide data allowed a better discriminant informative content than experimental mass spectra (overall accuracy higher than 87% in both collection 1 and 2). These results indicate that sequencing of peptides and proteins reduces the experimental noise affecting the raw mass spectra, and allows the extraction of more informative features available for the effective classification of samples. In addition, proteins and peptides features selected by SVM matched for 80% with the differentially expressed proteins identified by the MAProMa software.

Conclusions

These findings confirm the availability of the most label-free quantitative methods based on processing of spectral count and SEQUEST-based SCORE values. On the other hand, it stresses the usefulness of MudPIT data for a correct grouping of sample phenotypes, by applying both supervised and unsupervised learning algorithms. This capacity permit the evaluation of actual samples and it is a good starting point to translate proteomic methodology to clinical application.