TargetSearch - a Bioconductor package for the efficient preprocessing of GC-MS metabolite profiling data

Background  

Metabolite profiling, the simultaneous quantification of multiple metabolites in an experiment, is becoming increasingly popular, particularly with the rise of systems-level biology. The workhorse in this field is gas-chromatography hyphenated with mass spectrometry (GC-MS). The high-throughput of this technology coupled with a demand for large experiments has led to data pre-processing, i.e. the quantification of metabolites across samples, becoming a major bottleneck. Existing software has several limitations, including restricted maximum sample size, systematic errors and low flexibility. However, the biggest limitation is that the resulting data usually require extensive hand-curation, which is subjective and can typically take several days to weeks.     

Metabolite extraction from suspension-cultured mammalian cells for global metabolite profiling

Metabolite profiling of industrially important suspension-cultured mammalian cells is being increasingly used for rational improvement of bioprocesses. This requires the generation of global metabolite profiles that cover a broad range of metabolites and that are representative of the cells at the time of sampling. The protocol described here is a validated method for recovery of physiologically relevant amounts of key metabolites from suspension-cultured mammalian cells. The method is a two-step process consisting of initial quenching of the cells (to stop cellular metabolism and allow isolation of the cells) followed by extraction of the metabolites. The cells are quenched in 60% methanol supplemented with 0.85% (wt/vol) ammonium bicarbonate at -40 °C. Metabolites are then extracted from the quenched cells using two 100% methanol extractions followed by a single water extraction. Metabolite samples generated using this protocol are amenable to analysis by mass spectrometry-based techniques (e.g., gas chromatography-mass spectrometry, liquid chromatography-mass spectrometry), NMR spectroscopy and enzymatic assays.     

Development and application of an in-house reverse hybridization method for Chlamydia trachomatis genotyping

Aim

To develop and evaluate an in‐house reverse hybridization technique for Chlamydia trachomatis genotype identification.

Methods and Results

The evaluation of the developed and optimized reverse hybridization method on reference strains showed the specific detection of all genotypes. This technique showed its ability to type one inclusion‐forming unit of C. trachomatis genotype E and equivalent sensitivity to the Cobas TaqMan assay. It was also able to detect mixed infections in vitro. Application of the reverse hybridization method on 38 isolated C. trachomatis strains and their respective swabs allowed the detection of six urogenital genotypes D, E, F, G, H and K and one trachoma genotype B. Genotype E was the most prevalent, detected in 73% of the swab samples. Mixed infections were detected in 26% of swab cases.

Conclusion

The reverse hybridization technique is simple and does not require specialized instruments. It is powerful in the diagnosis of mixed infections and is suitable for use in epidemiological studies.

Significance and Impact of the Study

This technique allowed rapid C. trachomatis genotype identification.     

Intracellular metabolite profiling and the evaluation of metabolite extraction solvents for Clostridium carboxidivorans fermenting carbon monoxide

Clostridium carboxidivorans ferments CO, CO₂, and H₂ via the Wood-Ljungdahl pathway. CO, CO_2, and H₂ are unique substrates, unlike other carbon sources like glucose, so it is necessary to analyze intracellular metabolite profiles for gas fermentation by C. carboxidivorans for metabolic engineering. Moreover, it is necessary to optimize the metabolite extraction solvent specifically for C. carboxidivorans fermenting syngas. In comparison with glucose media, the gas media allowed significant abundance changes of 38 and 34 metabolites in the exponential and stationary phases, respectively. Especially, C. carboxidivorans cultivated in the gas media showed changes of fatty acid metabolism and higher levels of intracellular fatty acid synthesis possibly due to cofactor imbalance and slow metabolism. Meanwhile, the evaluation of extraction solvents revealed the mixture of water-isopropanol-methanol (2:2:5, v/v/v) to be the best extraction solvent, which showed a higher extraction capability and reproducibility than pure methanol, the conventional extraction solvent. This is the first metabolomic study to demonstrate the unique intracellular metabolite profiles of the gas fermentation compared to glucose fermentation, and to evaluate water-isopropanol-methanol as the optimal metabolite extraction solvent for C. carboxidivorans on gas fermentation.     

Evaluation of extraction processes for intracellular metabolite profiling of mammalian cells: matching extraction approaches to cell type and metabolite targets

In this study we report on the optimisation of the technologies for generation of a global metabolomics profile for intracellular metabolites in Chinese hamster ovary (CHO) cells. We evaluated the effectiveness of a range of different extraction methods applied to CHO cells which had been quenched using a previously optimised approach. The extraction methods tested included cold methanol, hot ethanol, acid, alkali and methanol/chloroform plus combinations of these. The extraction of metabolites using two 100% methanol extractions followed by a final water extraction recovered the largest range of metabolites. For the majority of metabolites, extracts generated in this manner exhibited the greatest recovery with high reproducibility. Therefore, this was the best extraction method for attaining a global metabolic profile from a single sample. However, another parallel extraction method (e.g. alkali) may also be required to maximise the range of metabolites recovered (e.g. non-polar metabolites).     

TagFinder for the quantitative analysis of gas chromatography--mass spectrometry (GC-MS)-based metabolite profiling experiments

MOTIVATION: Typical GC-MS-based metabolite profiling experiments may comprise hundreds of chromatogram files, which each contain up to 1000 mass spectral tags (MSTs). MSTs are the characteristic patterns of approximately 25-250 fragment ions and respective isotopomers, which are generated after gas chromatography (GC) by electron impact ionization (EI) of the separated chemical molecules. These fragment ions are subsequently detected by time-of-flight (TOF) mass spectrometry (MS). MSTs of profiling experiments are typically reported as a list of ions, which are characterized by mass, chromatographic retention index (RI) or retention time (RT), and arbitrary abundance. The first two parameters allow the identification, the later the quantification of the represented chemical compounds. Many software tools have been reported for the pre-processing, the so-called curve resolution and deconvolution, of GC-(EI-TOF)-MS files. Pre-processing tools generate numerical data matrices, which contain all aligned MSTs and samples of an experiment. This process, however, is error prone mainly due to (i) the imprecise RI or RT alignment of MSTs and (ii) the high complexity of biological samples. This complexity causes co-elution of compounds and as a consequence non-selective, in other words impure MSTs. The selection and validation of optimal fragment ions for the specific and selective quantification of simultaneously eluting compounds is, therefore, mandatory. Currently validation is performed in most laboratories under human supervision. So far no software tool supports the non-targeted and user-independent quality assessment of the data matrices prior to statistical analysis. TagFinder may fill this gap. Strategy: TagFinder facilitates the analysis of all fragment ions, which are observed in GC-(EI-TOF)-MS profiling experiments. The non-targeted approach allows the discovery of novel and unexpected compounds. In addition, mass isotopomer resolution is maintained by TagFinder processing. This feature is essential for metabolic flux analyses and highly useful, but not required for metabolite profiling. Whenever possible, TagFinder gives precedence to chemical means of standardization, for example, the use of internal reference compounds for retention time calibration or quantitative standardization. In addition, external standardization is supported for both compound identification and calibration. The workflow of TagFinder comprises, (i) the import of fragment ion data, namely mass, time and arbitrary abundance (intensity), from a chromatography file interchange format or from peak lists provided by other chromatogram pre-processing software, (ii) the annotation of sample information and grouping of samples into classes, (iii) the RI calculation, (iv) the binning of observed fragment ions of equal mass from different chromatograms into RI windows, (v) the combination of these bins, so-called mass tags, into time groups of co-eluting fragment ions, (vi) the test of time groups for intensity correlated mass tags, (vii) the data matrix generation and (viii) the extraction of selective mass tags supported by compound identification. Thus, TagFinder supports both non-targeted fingerprinting analyses and metabolite targeted profiling. AVAILABILITY: Exemplary TagFinder workspaces and test data sets are made available upon request to the contact authors. TagFinder is made freely available for academic use from http://www-en.mpimp-golm.mpg.de/03-research/researchGroups/01-dept1/Root_Metabolism/smp/TagFinder/index.html.     

1H NMR based metabolite profiling for optimizing the ethanol extraction of Wolfiporia cocos

Metabolite profiling of Wolfiporia cocos (family: Polyporaceae) had been much advancement in recent days, and its analysis by nuclear magnetic resonance (NMR) spectroscopy has become well established. However, the highly important trait of W. cocos still needs advanced protocols despite some standardization. Partial least squares discriminant analysis (PLS-DA) was used as the multivariate statistical analysis of the ¹H NMR data set. The PLS-DA model was validated, and the key metabolites contributing to the separation in the score plots of different ethanol W. cocos extract. ¹H NMR spectroscopy of W. cocos identified 33 chemically diverse metabolites in D₂O, consisting of 13 amino acids, 11 organic acids 2 sugars, 3 sugar alcohols, 1 nucleoside, and 3 others. Among these metabolites, the levels of tyrosine, proline, methionine, sarcosine, choline, acetoacetate, citrate, 4-aminobutyrate, aspartate, maltose, malate, lysine, xylitol, lactate threonine, leucine, valine, isoleucine, uridine, guanidoacetate, arabitol, mannitol, glucose, and betaine were increased in the 95% ethanol extraction sample compared with the levels in other samples, whereas level of acetate, phenylalanine, alanine, succinate, and fumarate were significantly increased in the 0% ethanol extraction sample. A biological triterpenoid, namely pachymic acid, was detected from different ethanol P. cocos extract using ¹H-NMR spectra were found in CDCl₃. This is the first report to perform the metabolomics profiling of different ethanol W. cocos extract. These researches suggest that W. cocos can be used to obtain substantial amounts of bioactive ingredients for use as potential pharmacological and nutraceuticals agents.     

program for reduction of data from community-level physiological profiling using Biolog microplates: rationale and critical interpretation of data

A program is offered that reduces data resulting from mixed-species inoculations into Biolog microplates. The procedures of the program are supported by a critical review of the literature relating to Biolog data reduction. The availability of standardized, accelerated data reduction protocols will facilitate study comparisons and allow efficient evaluation of new data reduction approaches.     

BASIC program for reduction of data from community-level physiological profiling using biolog microplates: rationale and critical interpretation of data

A BASIC program is offered that reduces data resulting from mixed-species inoculations into Biolog microplates. The procedures of the program are supported by a critical review of the literature relating to Biolog data reduction. The availability of standardized, accelerated data reduction protocols will facilitate study comparisons and allow efficient evaluation of new data reduction approaches.     

Breaking out of the biomed box: an audit assessment and recommendations for an in-house biomedical engineering program

In order to assess the current performance and to identify future growth opportunities of an in-house biomedical engineering (BME) program, senior management of Lehigh Valley Hospital (Allentown, Penn) engaged (in July 2001) the services of a clinical engineering consultant. Although the current in-house program was both functionally and financially sound, an independent audit had not been performed in over 4 years, and there were growing concerns by the BME staff related to the department's future leadership and long-term support from senior management. After an initial 2-month audit of the existing program, the consultant presented 41 separate recommendations for management's consideration. In order to refine and implement these recommendations, 5 separate committees were established to further evaluate a consolidated version of them, with the consultant acting as the facilitator for each group. Outcomes from each of the committees were used in the development of a formal business plan, which, upon full implementation, would not only strengthen and refine the current in-house service model but could also result in a substantial 3-year cost savings for the organization ($1,100,000 from existing operations, $500,000 in cost avoidance by in-sourcing postwarranty support of future capital equipment acquisitions). Another key outcome of the project was related to the development of a new master policy, titled the "Medical Equipment Management Program," complete with a newly defined state-of-the-art equipment scheduled inspection frequency model.     

Development of a new method for simultaneous extraction of chlorophylls and carotenoids from microalgal biomass

Journal of Applied Phycology - There are still limitations in the pigment extraction methods used in microalgae biomass, especially for laboratory scale. This work aimed to develop a simple method...     

Towards quantitative metabolomics of mammalian cells: Development of a metabolite extraction protocol

Metabolomics aims to quantify all metabolites within an organism, thereby providing valuable insight into the metabolism of cells. To study intracellular metabolites, they are first extracted from the cells. The ideal extraction procedure should immediately quench metabolism and quantitatively extract all metabolites, a significant challenge given the rapid turnover and physicochemical diversity of intracellular metabolites. We have evaluated several quenching and extraction solutions for their suitability for mammalian cells grown in suspension. Quenching with 60% methanol (buffered or unbuffered) resulted in leakage of intracellular metabolites from the cells. In contrast, quenching with cold isotonic saline (0.9% [w/v] NaCl, 0.5 °C) did not damage cells and effectively halted conversion of ATP to ADP and AMP, indicative of metabolic arrest. Of the 12 different extraction methods tested, cold extraction in 50% aqueous acetonitrile was superior to other methods. The recovery of a mixture of standards was excellent, and the concentration of extracted intracellular metabolites was higher than for the other methods tested. The final protocol is easy to implement and can be used to study the intracellular metabolomes of mammalian cells.     

A semiparametric empirical likelihood method for data from an outcome-dependent sampling scheme with a continuous outcome

Outcome-dependent sampling (ODS) schemes can be a cost effective way to enhance study efficiency. The case-control design has been widely used in epidemiologic studies. However, when the outcome is measured on a continuous scale, dichotomizing the outcome could lead to a loss of efficiency. Recent epidemiologic studies have used ODS sampling schemes where, in addition to an overall random sample, there are also a number of supplemental samples that are collected based on a continuous outcome variable. We consider a semiparametric empirical likelihood inference procedure in which the underlying distribution of covariates is treated as a nuisance parameter and is left unspecified. The proposed estimator has asymptotic normality properties. The likelihood ratio statistic using the semiparametric empirical likelihood function has Wilks-type properties in that, under the null, it follows a chi-square distribution asymptotically and is independent of the nuisance parameters. Our simulation results indicate that, for data obtained using an ODS design, the semiparametric empirical likelihood estimator is more efficient than conditional likelihood and probability weighted pseudolikelihood estimators and that ODS designs (along with the proposed estimator) can produce more efficient estimates than simple random sample designs of the same size. We apply the proposed method to analyze a data set from the Collaborative Perinatal Project (CPP), an ongoing environmental epidemiologic study, to assess the relationship between maternal polychlorinated biphenyl (PCB) level and children's IQ test performance.     

Development of a method for reconstruction of crowded NMR spectra from undersampled time-domain data

NMR is a unique methodology for obtaining information about the conformational dynamics of proteins in heterogeneous biomolecular systems. In various NMR methods, such as transferred cross-saturation, relaxation dispersion, and paramagnetic relaxation enhancement experiments, fast determination of the signal intensity ratios in the NMR spectra with high accuracy is required for analyses of targets with low yields and stabilities. However, conventional methods for the reconstruction of spectra from undersampled time-domain data, such as linear prediction, spectroscopy with integration of frequency and time domain, and analysis of Fourier, and compressed sensing were not effective for the accurate determination of the signal intensity ratios of the crowded two-dimensional spectra of proteins. Here, we developed an NMR spectra reconstruction method, “conservation of experimental data in analysis of Fourier” (Co-ANAFOR), to reconstruct the crowded spectra from the undersampled time-domain data. The number of sampling points required for the transferred cross-saturation experiments between membrane proteins, photosystem I and cytochrome b ₆ f, and their ligand, plastocyanin, with Co-ANAFOR was half of that needed for linear prediction, and the peak height reduction ratios of the spectra reconstructed from truncated time-domain data by Co-ANAFOR were more accurate than those reconstructed from non-uniformly sampled data by compressed sensing.     

Influence of ethanol concentration of extraction solvent on metabolite profiling for Salviae Miltiorrhizae Radix et Rhizoma extract by 1H NMR spectroscopy and multivariate data analysis

Salviae Miltiorrhizae Radix et Rhizoma (Danshen in China) and its related preparations are widely used in clinical practice due to its high medicinal value. In recent years, ¹H NMR technology has made great progress and demonstrated its unique advantages in the field of botanical metabolomics. In this study, ¹H NMR-based metabolomics was used to investigate the dissolution of various metabolites in Danshen as a function of ethanol concentration. ¹H NMR spectroscopy of Danshen extract identified 28 metabolites including 6 sugars, 11 amino acids, 3 organic acids, 4 salvianolic acids, and 4 tanshinones. Multivariate statistical analysis was used to classify and compare various Danshen extracts. PCA and HCA were used to obtain a global overview of the similarity in the samples and two-class OPLS-DA models were established for identifying characteristic metabolites. Then, ¹H-qNMR method was used to estimate the concentration of 22 metabolites, which is helpful to further describe the changes in metabolite ratios of various Danshen extracts. The result of this study laid the foundation for further biological activity research, and also provided an important reference for subsequent process research and quality control of Danshen related preparations.     

aTRAM - automated target restricted assembly method: a fast method for assembling loci across divergent taxa from next-generation sequencing data

Background

Assembling genes from next-generation sequencing data is not only time consuming but computationally difficult, particularly for taxa without a closely related reference genome. Assembling even a draft genome using de novo approaches can take days, even on a powerful computer, and these assemblies typically require data from a variety of genomic libraries. Here we describe software that will alleviate these issues by rapidly assembling genes from distantly related taxa using a single library of paired-end reads: aTRAM, automated Target Restricted Assembly Method. The aTRAM pipeline uses a reference sequence, BLAST, and an iterative approach to target and locally assemble the genes of interest.

Results

Our results demonstrate that aTRAM rapidly assembles genes across distantly related taxa. In comparative tests with a closely related taxon, aTRAM assembled the same sequence as reference-based and de novo approaches taking on average < 1 min per gene. As a test case with divergent sequences, we assembled >1,000 genes from six taxa ranging from 25 – 110 million years divergent from the reference taxon. The gene recovery was between 97 – 99% from each taxon.

Conclusions

aTRAM can quickly assemble genes across distantly-related taxa, obviating the need for draft genome assembly of all taxa of interest. Because aTRAM uses a targeted approach, loci can be assembled in minutes depending on the size of the target. Our results suggest that this software will be useful in rapidly assembling genes for phylogenomic projects covering a wide taxonomic range, as well as other applications. The software is freely available http://www.github.com/juliema/aTRAM.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-015-0515-2) contains supplementary material, which is available to authorized users.     

Experiences with an interactive computer program for recording biometrical health screening data

The nurses measurements of biometrical health screening data comprise the central elements in our multiphasic preventitive medical investigation in Malmö. It is also vitally importantfor the formation processing, evaluation and actions of the screening that all test values and other data can be supplied to a uniform computer bank without delay in terms of secondary transferral routines. In our ongoing population investigation project in Malmö, this has been accomplished by on-line computer-programs for the different components of the screening. Here we describe the practical experiences with the subroutine used by the nurses to feed the biometrical test results to the computer bank.     

Principal component analysis of hormone profiling data suggests an important role for cytokinins in regulating leaf growth and senescence of salinized tomato

High throughput analytical methods allow phytohormonal profiling, but the magnitude of the data generated makes it difficult to draw firm conclusions about the physiological roles of different compounds. Principal component analysis (PCA) was used as a mathematical tool to evaluate relationships between physiological and hormonal variables in two experiments with salinised tomato. When tomato plants (cv Boludo F1) were grafted onto a recombinant inbred line (RIL) population derived from a Solanum lycopersicum x S. cheesmaniae cross and grown under moderate salinity (75 mM NaCl) for 100 days under greenhouse conditions, PCA revealed an important role for leaf xylem cytokinins (CKs) in controlling leaf growth and photosystem II efficiency (Fv/Fm) and thus crop productivity under salinity. PCA analysis from a similar experiment, with ungrafted tomato grown under highly saline (100 mM NaCl) conditions, that evaluated the temporal sequence of leaf growth (as relative growth rate, LRGR) and senescence and hormone concentrations, revealed a similar influence of CKs on both processes, since Fv/Fm and LRGR were strongly loaded along the two principal components and placed in the same cluster as leaf trans-zeatin and/or related to other CK-related parameters. The conservative behaviour of the eigen vectors for Fv/Fm and the analyzed phytohormones in different compartments (xylem, leaf and root) between different experiments suggests an important role for CKs in regulating leaf senescence, while CKs and other hormones seem to regulate leaf growth under salinity.Key words: cytokinins, leaf growth, principal component analysis, salinity, senescence, tomatoAn important paradigm of plant growth regulation is that plant roots can sense their environment, alter their metabolism and transmit chemical signals via the xylem to the shoots to regulate shoot physiology.¹ Much work has aimed to substantiate this “chemical signaling hypothesis” by determining the production and distribution of various signals such as the plant hormones ABA, cytokinins, the ethylene precursor ACC and various nutrient ions.² Although this work has largely been “ABA-centric”, in part due to its relative ease of measurement, the advent of high-throughput, multi-analyte physicochemical techniques to quantify plant hormones^3,4 greatly amplifies the information available from analyses of long-distance signaling, and allows us to move away from a priori assumptions as to which hormone(s) might be physiologically relevant to particular processes. Ultimately, interpreting this information is necessary to provide a sound physiological basis to underpin efforts aimed at manipulating long-distance signaling in planta.Full spectrum hormone profiling can potentially assay more variables per sample than the typical number of samples assayed. Principal component analysis (PCA) is a mathematical algorithm that reduces the dimensionality of the data set while retaining most of the inherent variation.⁵ This is achieved by identifying directions, called principal components, along which the variation in the data is maximal. By using few components, each sample can be represented by relatively few numbers instead of values for many variables.⁶ PCA identifies new variables, the principal components, which are linear combinations of the original variables, and may be an appropriate technique to aid understanding of hormone profiling experiments.Recently, we grew a commercial tomato cultivar (cv Boludo F1) grafted onto 100 rootstocks from a population of recombinant inbred lines derived from a Solanum lycopersicum x S. cheesmaniae cross, and exposed the plants to moderate (75 mM NaCl) salinity for 100 days.⁷ The rootstock generated considerable variability in vegetative vigour (assessed as fresh weight of a fully expanded leaf, LFW) and leaf senescence (assessed by the chlorophyll fluorescence parameter Fv/Fm in that leaf). Ionic and hormonal factor(s) putatively regulating these processes (xylem concentrations of Na⁺ and K⁺ and ABA; the cytokinins trans-zeatin, Z, and the storage form trans-zeatin riboside, ZR; and the ethylene precursor 1-aminocyclopropane-1-carboxylic acid, ACC) were analyzed in leaf xylem sap collected 50 days after salinisation in seven graft combinations of contrasting vigour. Since different xylem parameters showed a high degree of autocorrelation, PCA was performed in order to gain further insights about their real contribution to the physiological processes. Xylem K⁺, K⁺/Na⁺, the active cytokinins Z and ZR, its sum (Z + ZR) and ratio (Z/ZR), and especially the ratio between cytokinins and ACC (Z (ZR)/ACC and Z + ZR/ACC) were strong and positively loaded into the first principal component (PC1) determining both LFW and Fv/Fm (Fig. 1A⁷). Although other variables are included in the same cluster at the 95% of confidence level, their strength in PC1 was much weaker (e.g., Na⁺, Z/ABA, ACC). Does this PCA output provide generic information about the hormonal processes regulating leaf growth and Fv/Fm of salinized plants, or is it specific to the particular methodological conditions (duration of salinization, choice of genotypes or plant compartment) imposed by our experimental design?Open in a separate windowFigure 1Two axes of a principal components (PC1, PC2) analysis showing plant productivity trait vectors (leaf fresh weight, LFW or relative growth rate, LRGR, and Fv/Fm, indicated by arrows) and the position of various hormonal and ionic variables (denoted by abbreviations) for the long-term grafting experiment which analysed xylem variables (A), and the medium-term experiments with ungrafted plants that analyzed leaf (B) and root (C) variables. Arrows indicate eigen vectors representing the strength (given by the length of the vector) and direction of the trait correlation relative to the first two principal components (PC1, PC2). The circles enclose those variables that fall into the same cluster (95% confidence level). Abbreviations for the variables are given thus: ABA, abscisic acid; AC , 1-aminocyclopropane-1-carboxylic acid; IAA , indole-3-acetic acid; Z, trans-zeatin; ZR, trans-zeatin riboside.To answer this question, PCA was performed on data obtained from a similar experiment where a single genotype of tomato (cv Moneymaker) was exposed to high (100 mM NaCl) salinity for 22 days under hydroponic conditions in a controlled environment chamber.^8,9 Although the same ionic and hormonal variables (including the auxin indole-3-acetic acid, IAA) were assayed in both roots and leaves, xylem ion concentrations were not quantified. Therefore PCA was conducted using leaf relative growth rate (LRGR) as an indicator of vegetative vigour (due to the temporal variation of leaf fresh weight in young plants) and senescence (Fv/Fm), and hormonal and ionic variables measured in leaves (Fig. 1B) and roots (Fig. 1C).Both physiological variables (LRGR and Fv/Fm) were significantly loaded into the two major principal components (PC1 and PC2) explaining more than 90% of the variance, and in which most of the leaf and root ionic and hormonal parameters were strongly associated in three clusters (enclosed within the circles, Fig. 1B and C). As in the PCA from the grafting experiment (Fig. 1A), LRGR and Fv/Fm, were also placed in the same cluster as the cytokinin Z and the ionic variables K⁺ and K⁺/Na⁺ in the leaf (Fig. 1B), although some dissociation between them was observed, probably due to the temporal dynamics of these variables in the same plant organ. Moreover, most of the other CK-related variables (ZR, Z + ZR, Z + ZR/ACC, Z + ZR/ABA) were associated in a distinct cluster that was also strongly loaded along with Fv/Fm parameter in PC1 (Fig. 1B). When the root data were considered (Fig. 1C), all these CK-related (with the exception of Z) and ionic variables were placed in the same cluster as Fv/Fm and loaded into PC1 explaining 71% of variance.Interestingly, the position and value of the eigen vector defining Fv/Fm remained highly conservative and positively associated with hormonal (ZR, Z + ZR, Z + ZR/ACC) and ionic (K⁺, K⁺/Na⁺) variables independent of the experimental design and organ analyzed. However, the position of the vector for LRGR was more variable since its relationship with PC2 and some hormonal parameters (e.g., IAA, IAA/Z + ZR) was positive in the leaves and negative in the roots. Particular attention should be paid to leaf Z concentration since it was highly related to both Fv/Fm and leaf growth vectors in both experiments (Fig. 1A and B), while this hormone was placed in an opposite (to Fv/Fm) or orthogonal (to LRGR) vector in the roots (Fig. 1C). Only two of the hormones analyzed showed differential responses to salinity in leaves and roots suggesting an important role in biomass partitioning: Z and IAA decreased in leaves and increased in roots.⁸ Similar responses of both roots and leaves (for the other hormones) may indicate the importance of root-to-shoot signaling in maintaining hormone homeostasis in particular organs and integrating the whole plant response to the stress, but reciprocal grafting experiments with hormone-content or -sensitivity mutants are required to test this hypothesis for each compound.¹⁰ Additionally, PCA also revealed a conserved position for some hormonal parameters such as ACC and ABA tissue concentrations (Fig. 1B and C) and also the ratio ACC/ABA in both tissues and leaf xylem (Fig. 1A–C), which were always placed in clusters opposed to Fv/Fm (ACC, ABA) and leaf growth (ACC/ABA), supporting a role of these hormones in negatively regulating salt-induced leaf senescence⁹ and growth.⁷The relatively conservative geometric positions of some ionic and hormonal variables, and their co-occurrence with physiological variables of interest, suggest that PCA was relatively insensitive to methodological issues such as duration of salinisation, choice of genotypes or plant compartment. Moreover, similar PCAs for both tissue (Fig. 1B and C) and xylem (Fig. 1A) samples indicate the adequacy of the latter (with its considerably decreased sample preparation time) to infer physiological relationships. Particularly noteworthy was the clustering of cytokinin-related variables with the chlorophyll fluorescence parameter Fv/Fm in both experiments and organs, as well as the leaf Z concentration and growth, which highlighted the potential physiological importance of cytokinins in a way that was not so apparent from inspection of a typical correlation matrix which analyses two variables at a time. Accordingly, we have instigated experiments manipulating the cytokinin status of salinised plants by selectively overexpressing the ipt gene. That these plants grew better, and showed delayed senescence, under salinity (Ghanem ME, unpublished results) supports our use of PCA as a tool to explore temporal^8,9 and genetic⁷ correlations.