A Non-parametric Cutout Index for Robust Evaluation of Identified Proteins

This paper proposes a novel, automated method for evaluating sets of proteins identified using mass spectrometry. The remaining peptide-spectrum match score distributions of protein sets are compared to an empirical absent peptide-spectrum match score distribution, and a Bayesian non-parametric method reminiscent of the Dirichlet process is presented to accurately perform this comparison. Thus, for a given protein set, the process computes the likelihood that the proteins identified are correctly identified. First, the method is used to evaluate protein sets chosen using different protein-level false discovery rate (FDR) thresholds, assigning each protein set a likelihood. The protein set assigned the highest likelihood is used to choose a non-arbitrary protein-level FDR threshold. Because the method can be used to evaluate any protein identification strategy (and is not limited to mere comparisons of different FDR thresholds), we subsequently use the method to compare and evaluate multiple simple methods for merging peptide evidence over replicate experiments. The general statistical approach can be applied to other types of data (e.g. RNA sequencing) and generalizes to multivariate problems.Mass spectrometry is the predominant tool for characterizing complex protein mixtures. Using mass spectrometry, a heterogeneous protein sample is digested into peptides, which are separated by various features (e.g. retention time and mass-to-charge ratio), and fragmented to produce a large collection of spectra; these fragmentation spectra are matched to peptide sequences, and the peptide-spectrum matches (PSMs)¹ are scored (1). PSM scores from different peptide search engines and replicate experiments can be assembled to produce consensus scores for each peptide (2, 3). These peptide search results are then used to identify proteins (4).Inferring the protein content from these fragment ion spectra is difficult, and statistical methods have been developed with that goal. Protein identification methods (5–8) rank proteins according to the probability of their being present in the sample. Complementary target-decoy methods evaluate the proteins identified by searching fragmentation spectra against proteins that might be present (targets) and proteins that are absent (decoys). An identified target protein counts as a correct identification (increasing the estimated sensitivity), whereas each identified decoy protein counts as an incorrect identification (lowering the estimated specificity).Current target-decoy methods estimate the protein-level false discovery rate (FDR) for a set of identified proteins (9, 10), as well as the sensitivity at a particular arbitrary FDR threshold (11); however, these methods have two main shortcomings.First, current methods introduce strong statistical biases, which can be conservative (10) or optimistic (12) in different settings. These biases make current approaches unreliable for comparing different identification methods, because they implicitly favor methods that use similar assumptions. Automated evaluation tools that can be run without user-defined parameters are necessary in order to compare and improve existing analysis tools (13).Second, existing evaluation methods do not produce a single quality measure; instead, they estimate both FDR and sensitivity (which is estimated using the “absolute sensitivity,” which treats all targets as present and counts them as true identifications). For data sets with known protein contents (e.g. the protein standard data set considered), the absolute sensitivity is estimable; however, for more complex data sets with unknown contents, the measurement indicates the relative sensitivity. Even if one ignores statistical biases, there is currently no method for choosing a non-arbitrary FDR threshold, and it is currently not possible to decide which protein set is superior—one with a lower sensitivity and stricter FDR, or another with a higher sensitivity and less stringent FDR. The former is currently favored but might result in significant information loss. Arbitrary thresholds have significant effects: in the yeast data analyzed, 1% and 5% FDR thresholds, respectively, yielded 1289 and 1570 identified protein groups (grouping is discussed in the supplementary “Methods” section). Even with such a simple data set, this subtle change results in 281 more target identifications, of which unknown subsets of 66 (0.05 × 1570 − 0.01 × 1289 ≈ 66) are expected to be false identifications and 215 are expected to be true identifications (281 − 66 = 215).Here we introduce the non-parametric cutout index (npCI), a novel, automated target-decoy method that can be used to compute a single robust and parameter-free quality measure for protein identifications. Our method does not require prior expertise in order for the user to select parameters or run the computation. The npCI employs target-decoy analysis at the PSM level, where its assumptions are more applicable (4). Rather than use assumptions to model PSM scores matching present proteins, our method remains agnostic to the characteristics of present proteins and analyzes PSMs not explained by the identified proteins. If the correct present set of proteins is known, then the distribution of remaining, unexplained PSM scores resembles the decoy distribution (14). We extend this idea and present a general graphical framework to evaluate a set of protein identifications by computing the likelihood that the remaining PSMs and decoy PSMs are drawn from the same distribution (Fig. 1).Open in a separate windowFig. 1.Schematic for non-parametric probabilistic evaluation of identified proteins. Under the supposition that the identified protein set (blue) is present, all peptides matching those proteins (also blue) might be present and have an unknown score distribution. When the correct set of proteins is identified, the remaining peptides (i.e. those not matching any shaded proteins in this figure) have a score distribution resembling that of absent peptides. Thus, the similarity of the remaining peptide score distribution (red dashed line) to the absent peptide score distribution (black solid line) determines the quality of the identified proteins.Existing non-parametric statistical tests evaluating the similarity between two collections of samples (e.g. Kolmogorov–Smirnov test, used in Ref. 14, and the Wilcoxon signed rank test) were inadequate because infrequent but significant outliers (e.g. high-scoring PSMs) are largely ignored by these methods. Likewise, information-theoretic measures, such as the Kullback–Leibler divergence, were inadequate because they require a prior on the smoothing parameter that weighs more smoothing and higher similarity against less smoothing and lower similarity (a problem reminiscent of the original compromise between sensitivity and FDR); without the application of such a prior, the optimal Kullback–Leibler divergence occurs with infinite smoothing, which will make any distributions equal, rendering them completely uninformative and thus making it impossible to distinguish one identified protein set from another. For these reasons, we derived a novel, Bayesian, non-parametric process to compute the likelihood that two continuous collections are drawn from the same distribution. It can be used to provide a robust and efficient evaluation of discoveries.     

Multiple Protein Analysis of Formalin-fixed and Paraffin-embedded Tissue Samples with Reverse phase Protein Arrays

Reverse-phase protein arrays (RPPAs) have become an important tool for the sensitive and high-throughput detection of proteins from minute amounts of lysates from cell lines and cryopreserved tissue. The current standard method for tissue preservation in almost all hospitals worldwide is formalin fixation and paraffin embedding, and it would be highly desirable if RPPA could also be applied to formalin-fixed and paraffin embedded (FFPE) tissue. We investigated whether the analysis of FFPE tissue lysates with RPPA would result in biologically meaningful data in two independent studies. In the first study on breast cancer samples, we assessed whether a human epidermal growth factor receptor (HER) 2 score based on immunohistochemistry (IHC) could be reproduced with RPPA. The results showed very good concordance between the IHC and RPPA classifications of HER2 expression. In the second study, we profiled FFPE tumor specimens from patients with adenocarcinoma and squamous cell carcinoma in order to find new markers for differentiating these two subtypes of non-small cell lung cancer. p21-activated kinase 2 could be identified as a new differentiation marker for squamous cell carcinoma. Overall, the results demonstrate the technical feasibility and the merits of RPPA for protein expression profiling in FFPE tissue lysates.Many diseases are characterized by the expression of specific proteins and the activation status of distinct signaling pathways (1). Thus, protein expression profiling and activation patterns are instrumental for understanding disease, the development of effective treatments, and the identification of patients who will respond to particular therapies. Traditional ways of analyzing protein expression (e.g. Western blot) can be used for these purposes but often are labor intensive, have low throughput, and consume high sample volumes. Reverse-phase protein array (RPPA)¹ technology is a very promising method that circumvents these issues (2–4). For RPPA, minute amounts of whole protein lysates from a multitude of samples are spotted onto slides, and individual proteins are detected via protein-specific antibodies. This enables medium- to high-throughput analysis of precious low-volume sample material.Lysates for RPPA have so far been generated mainly from cell lines or fresh frozen tissue. However, because of the high amount of effort involved in the use of liquid nitrogen for sample preservation, in almost all hospitals worldwide formalin fixation and paraffin embedding is the preferred method for tissue preservation. Therefore, it would be highly desirable if protein-specific epitopes could be quantitatively extracted and analyzed from formalin-fixed and paraffin embedded (FFPE) tissue, as this would make the majority of clinical specimens accessible for mechanistic protein-based research.In recent years, several research groups have established protocols for protein extraction from FFPE tissue. Common to all of them is the use of high concentrations of ionic detergents, such as sodium dodecyl sulfate, and high temperature. It was shown that these methods even make it possible to extract full-length proteins from FFPE tissue (5–12). The coefficient of variation of the relative extraction efficiency based on Western blot and densitometric assessment of actin typically is below 20% (13). To assess whether the analysis of FFPE tissue lysates would result in biologically meaningful data, we analyzed FFPE breast cancer tissue samples by RPPA for the expression of human epidermal growth factor receptor 2 (HER2) and compared it to HER2 assessment by the gold standard used in clinical practice, which is based on immunohistochemistry (IHC). Successful recovery of HER2 from FFPE tissue should result in concordant HER2 classification between RPPA and IHC.In the second part of the study, FFPE samples of non-small cell lung cancer (NSCLC) were examined via RPPA. Samples from two subtypes of NSCLC, adenocarcinoma (AC) and squamous cell carcinoma (SCC), were analyzed for more than 150 proteins, including two proteins that are known to be differentially expressed between the two subtypes. The objectives of this analysis were to further assess the validity of the approach by confirming the two positive controls and to identify new markers for the differentiation of the two subtypes of NSCLC.     

Identification of New PNEPs Indicates a Substantial Non-PEXEL Exportome and Underpins Common Features in Plasmodium falciparum Protein Export

Malaria blood stage parasites export a large number of proteins into their host erythrocyte to change it from a container of predominantly hemoglobin optimized for the transport of oxygen into a niche for parasite propagation. To understand this process, it is crucial to know which parasite proteins are exported into the host cell. This has been aided by the PEXEL/HT sequence, a five-residue motif found in many exported proteins, leading to the prediction of the exportome. However, several PEXEL/HT negative exported proteins (PNEPs) indicate that this exportome is incomplete and it remains unknown if and how many further PNEPs exist. Here we report the identification of new PNEPs in the most virulent malaria parasite Plasmodium falciparum. This includes proteins with a domain structure deviating from previously known PNEPs and indicates that PNEPs are not a rare exception. Unexpectedly, this included members of the MSP-7 related protein (MSRP) family, suggesting unanticipated functions of MSRPs. Analyzing regions mediating export of selected new PNEPs, we show that the first 20 amino acids of PNEPs without a classical N-terminal signal peptide are sufficient to promote export of a reporter, confirming the concept that this is a shared property of all PNEPs of this type. Moreover, we took advantage of newly found soluble PNEPs to show that this type of exported protein requires unfolding to move from the parasitophorous vacuole (PV) into the host cell. This indicates that soluble PNEPs, like PEXEL/HT proteins, are exported by translocation across the PV membrane (PVM), highlighting protein translocation in the parasite periphery as a general means in protein export of malaria parasites.     

Plant evolution: pulses of extinction and speciation in gymnosperm diversity

Living gymnosperms represent the survivors of ancient seed plant lineages whose fossil record reaches back 270 million years. Two recent studies find that recent pulses of extinction and speciation have shaped today's gymnosperm diversity, contradicting the widespread assumption that gymnosperms have remained largely unchanged for tens of millions of years.     

Is sirolimus a therapeutic option for patients with progressive pulmonary lymphangioleiomyomatosis?

Background

Lymphangioleiomyomatosis (LAM) is a rare lung disease characterised by progressive airflow obstruction. No effective medical treatment is available but therapy with sirolimus has shown some promise. The aim of this observational study was to evaluate sirolimus in progressive LAM.

Methods

Sirolimus (trough level 5 - 10 ng/ml) was administered to ten female patients (42.4 ± 11.9 years) with documented progression. Serial pulmonary function tests and six-minute-walk-distance (6-MWD) assessments were performed.

Results

The mean loss of FEV₁ was -2.30 ± 0.52 ml/day before therapy and a significant mean gain of FEV₁ of 1.19 ± 0.26 ml/day was detected during treatment (p = 0.001). Mean FEV₁ and FVC at baseline were 1.12 ± 0.15 l (36.1 ± 4.5%pred.) and 2.47 ± 0.25 l (69.2 ± 6.5%pred.), respectively. At three and six months during follow-up a significant increase of FEV₁ and FVC was demonstrated (3 months ΔFEV₁: 220 ± 82 ml, p = 0.024; 6 months ΔFEV₁: 345 ± 58 ml, p = 0.001); (3 months ΔFVC: 360 ± 141 ml, p = 0.031; 6 months ΔFVC: 488 ± 138 ml, p = 0.006). Sirolimus was discontinued in 3 patients because of serious recurrent lower respiratory tract infection or sirolimus-induced pneumonitis. No deaths and no pneumothoraces occurred during therapy.

Conclusions

Our data suggest that sirolimus might be considered as a therapeutic option in rapidly declining LAM patients. However, sirolimus administration may be associated with severe respiratory adverse events requiring treatment cessation in some patients. Moreover, discontinuation of sirolimus is mandatory prior to lung transplantation.     

Recommendations for mass spectrometry data quality metrics for open access data (corollary to the Amsterdam Principles)

Policies supporting the rapid and open sharing of proteomic data are being implemented by the leading journals in the field. The proteomics community is taking steps to ensure that data are made publicly accessible and are of high quality, a challenging task that requires the development and deployment of methods for measuring and documenting data quality metrics. On September 18, 2010, the United States National Cancer Institute convened the "International Workshop on Proteomic Data Quality Metrics" in Sydney, Australia, to identify and address issues facing the development and use of such methods for open access proteomics data. The stakeholders at the workshop enumerated the key principles underlying a framework for data quality assessment in mass spectrometry data that will meet the needs of the research community, journals, funding agencies, and data repositories. Attendees discussed and agreed up on two primary needs for the wide use of quality metrics: 1) an evolving list of comprehensive quality metrics and 2) standards accompanied by software analytics. Attendees stressed the importance of increased education and training programs to promote reliable protocols in proteomics. This workshop report explores the historic precedents, key discussions, and necessary next steps to enhance the quality of open access data. By agreement, this article is published simultaneously in the Journal of Proteome Research, Molecular and Cellular Proteomics, Proteomics, and Proteomics Clinical Applications as a public service to the research community. The peer review process was a coordinated effort conducted by a panel of referees selected by the journals.     

Generic ecological impact assessments of alien species in Norway: a semi-quantitative set of criteria

The ecological impact assessment scheme that has been developed to classify alien species in Norway is presented. The underlying set of criteria enables a generic and semi-quantitative impact assessment of alien species. The criteria produce a classification of alien species that is testable, transparent and easily adjustable to novel evidence or environmental change. This gives a high scientific and political legitimacy to the end product and enables an effective prioritization of management efforts, while at the same time paying attention to the precautionary principle. The criteria chosen are applicable to all species regardless of taxonomic position. This makes the assessment scheme comparable to the Red List criteria used to classify threatened species. The impact of alien species is expressed along two independent axes, one measuring invasion potential, the other ecological effects. Using this two-dimensional approach, the categorization captures the ecological impact of alien species, which is the product rather than the sum of spread and effect. Invasion potential is assessed using three criteria, including expected population lifetime and expansion rate. Ecological effects are evaluated using six criteria, including interactions with native species, changes in landscape types, and the potential to transmit genes or parasites. Effects on threatened species or landscape types receive greater weightings.     

The artificial humic substance HS1500 does not inhibit photosynthesis of the green alga <Emphasis Type="Italic">Desmodesmus armatus</Emphasis> in vivo but interacts with the photosynthetic apparatus of isolated spinach thylakoids in vitro

Humic substances (HSs) can influence the growth and composition of freshwater phytoplankton assemblage. Since HSs contain many phenolic and quinonic moieties and cause growth reductions in eco-physiological field experiments, HSs are considered photosystem II herbicides. To test this specific mode of action in vivo and in vitro, respectively, we used intact cells of the green alga Desmodesmus armatus, as well as thylakoids isolated from spinach (Spinacia oleracea) as a model system for the green algal chloroplast. Photosynthetic electron transport was measured as oxygen evolution and variable chlorophyll fluorescence. The in vivo effect of the artificial humic substance HS1500 on algae consisted of no impact on photosynthesis–irradiance curves of intact green algae compared to untreated controls. In contrast, addition of HS1500 to isolated thylakoids resulted in light-induced oxygen consumption (Mehler reaction) as an in vitro effect. Fluorescence induction kinetics of HS-treated thylakoids revealed a large static quenching effect of HS1500, but no inhibitory effect on electron transport. For the case of intact algal cells, we conclude that the highly hydrophilic and rather large molecules of HS1500 are not taken up in effective quantities and, therefore, cannot interfere with photosynthesis. The in vitro tests show that HS1500 has no inhibitory effect on photosystem II but operates as a weak, oxygen-consuming Hill acceptor at photosystem I. Hence, the results indicate that eco-physiological field experiments should focus more strongly on effects of HSs on extracellular features, such as reducing and red-shifting the underwater light field or influencing nutrient availability by cation exchange within the plankton network.     

Production of 1,3-propanediol by Clostridium butyricum DSM 5431 and product tolerant mutants in fedbatch culture: Feeding strategy for glycerol and ammonium

Summary A fedbatch strategy was developed coupling the feeding of the two inhibitory substrates glycerol and ammonium to alkali consumption. A continuous, automated substrate addition was achieved responding directly to the needs of the culture. Thus substrate concentrations were kept on a constant low, but non limiting level. The feeding was applied for the cultivation of Clostridium butyricum DSM 5431 and mutants with increased product tolerance. Compared to fedbatch cultivations with intermittent feeding cultivation times were considerably shortened.     

Elective frozen-thawed embryo transfer (FET) in women at risk for ovarian hyperstimulation syndrome

Elective cryopreservation of cultured embryos has become a treatment option for women at risk for ovarian hyperstimulation syndrome (OHSS). The aim of our study was to investigate the outcome of elective cryopreservation and consecutive frozen-thawed embryo transfer (FET) in a large IVF clinic in Austria. A total of 6104 controlled ovarian hyperstimulation cycles (COH) were performed on 2998 patients including 200 patients (6.7%) who were undergoing elective cryopreservation and FET due to high risk of OHSS. We estimated the cumulative live birth rate using the Kaplan-Meier method and evaluated independent predictors for successful live births with a Cox model. A total of 270 frozen-thawed embryo transfers were performed on 200 patients with up to 4 transfers per patient. The first embryo transfer showed a live birth rate of 42.0%, the second transfer showed a cumulative rate of 58.5%. After a total of 4 FETs from the same COH cycle, a cumulative live birth rate of 61.0% per COH cycle could be achieved. Four cases of OHSS occurred amongst these patients (2.0%), all of them of moderate severity. Multivariate analysis identified maternal age, the use of assisted hatching and the number of embryos transferred at the blastocyst stage as independent predictors for cumulative live birth. Our study clearly suggests that elective FET is safe and shows excellent cumulative live birth rates. This concept can, therefore, be used to avoid the severe adverse events caused by COH and the inefficient use of cultured embryos.