MatureP: prediction of secreted proteins with exclusive information from their mature regions

Automated machine learning for genome wide association studies

MOTIVATION

Genome-wide association studies (GWAS) present several computational and statistical challenges for their data analysis, including knowledge discovery, interpretability, and translation to clinical practice.

RESULTS

We develop, apply, and comparatively evaluate an automated machine learning (AutoML) approach, customized for genomic data that delivers reliable predictive and diagnostic models, the set of genetic variants that are important for predictions (called a biosignature), and an estimate of the out-of-sample predictive power. This AutoML approach discovers variants with higher predictive performance compared to standard GWAS methods, computes an individual risk prediction score, generalizes to new, unseen data, is shown to better differentiate causal variants from other highly correlated variants, and enhances knowledge discovery and interpretability by reporting multiple equivalent biosignatures.

AVAILABILITY AND IMPLEMENTATION

Code for this study is available at: https://github.com/mensxmachina/autoML-GWAS. JADBio offers a free version at: https://jadbio.com/sign-up/. SNP data can be downloaded from the EGA repository (https://ega-archive.org/). PRS data are found at: https://www.aicrowd.com/challenges/opensnp-height-prediction. Simulation data to study population structure can be found at: https://easygwas.ethz.ch/data/public/dataset/view/1/.

Evolutionary adaptation of the protein folding pathway for secretability

Secretory preproteins of the Sec pathway are targeted post-translationally and cross cellular membranes through translocases. During cytoplasmic transit, mature domains remain non-folded for translocase recognition/translocation. After translocation and signal peptide cleavage, mature domains fold to native states in the bacterial periplasm or traffic further. We sought the structural basis for delayed mature domain folding and how signal peptides regulate it. We compared how evolution diversified a periplasmic peptidyl-prolyl isomerase PpiA mature domain from its structural cytoplasmic PpiB twin. Global and local hydrogen-deuterium exchange mass spectrometry showed that PpiA is a slower folder. We defined at near-residue resolution hierarchical folding initiated by similar foldons in the twins, at different order and rates. PpiA folding is delayed by less hydrophobic native contacts, frustrated residues and a β-turn in the earliest foldon and by signal peptide-mediated disruption of foldon hierarchy. When selected PpiA residues and/or its signal peptide were grafted onto PpiB, they converted it into a slow folder with enhanced in vivo secretion. These structural adaptations in a secretory protein facilitate trafficking.

In planta RNAi approach targeting three M. incognita effector genes disturbed the process of infection and reduced plant susceptibility

Meloidogyne incognita is the most economically important species of the root-knot nematode complex causing damage to several crops worldwide. During parasitism in host plants, M. incognita secretes several effector proteins to suppress the plant immune system, manipulate the plant cell cycle, and promote parasitism. Several effector proteins have been identified, but their relationship with plant parasitism by M. incognita has not been fully confirmed. Herein, the Minc01696, Minc00344, and Minc00801 putative effector genes were evaluated to assess their importance during soybean and Nicotiana tabacum parasitism by M. incognita. For this study, we used in planta RNAi technology to overexpress dsRNA molecules capable of producing siRNAs that target and downregulate these nematode effector genes. Soybean composite roots and N. tabacum lines were successfully generated, and susceptibility level to M. incognita was evaluated. Consistently, both transgenic soybean roots and transgenic N. tabacum lines carrying the RNAi strategy showed reduced susceptibility to M. incognita. The number of galls per plant and the number of egg masses per plant were reduced by up to 85% in transgenic soybean roots, supported by the downregulation of effector genes in M. incognita during parasitism. Similarly, the number of galls per plant, the number of egg masses per plant, and the nematode reproduction factor were reduced by up to 83% in transgenic N. tabacum lines, which was also supported by the downregulation of the Minc00801 effector gene during parasitism. Therefore, our data indicate that all three effector genes can be a target in the development of new biotechnological tools based on the RNAi strategy in economically important crops for M. incognita control.

Just Add Data: automated predictive modeling for knowledge discovery and feature selection

Fully automated machine learning (AutoML) for predictive modeling is becoming a reality, giving rise to a whole new field. We present the basic ideas and principles of Just Add Data Bio (JADBio), an AutoML platform applicable to the low-sample, high-dimensional omics data that arise in translational medicine and bioinformatics applications. In addition to predictive and diagnostic models ready for clinical use, JADBio focuses on knowledge discovery by performing feature selection and identifying the corresponding biosignatures, i.e., minimal-size subsets of biomarkers that are jointly predictive of the outcome or phenotype of interest. It also returns a palette of useful information for interpretation, clinical use of the models, and decision making. JADBio is qualitatively and quantitatively compared against Hyper-Parameter Optimization Machine Learning libraries. Results show that in typical omics dataset analysis, JADBio manages to identify signatures comprising of just a handful of features while maintaining competitive predictive performance and accurate out-of-sample performance estimation.

Minc03328 effector gene downregulation severely affects Meloidogyne incognita parasitism in transgenic Arabidopsis thaliana

Minc03328 effector gene downregulation triggered by in planta RNAi strategy strongly reduced plant susceptibility to Meloidogyne incognita and suggests that Minc03328 gene is a promising target for the development of genetically engineered crops to improve plant tolerance to M. incognita. Meloidogyne incognita is the most economically important species of root-knot nematodes (RKN) and causes severe damage to crops worldwide. M. incognita secretes several effector proteins to suppress the host plant defense response, and manipulate the plant cell cycle and other plant processes facilitating its parasitism. Different secreted effector proteins have already been identified in M. incognita, but not all have been characterized or have had the confirmation of their involvement in nematode parasitism in their host plants. Herein, we characterized the Minc03328 (Minc3s00020g01299) effector gene, confirmed its higher expression in the early stages of M. incognita parasitism in plants, as well as the accumulation of the Minc03328 effector protein in subventral glands and its secretion. We also discuss the potential for simultaneous downregulation of its paralogue Minc3s00083g03984 gene. Using the in planta RNA interference strategy, Arabidopsis thaliana plants overexpressing double-stranded RNA (dsRNA) were generated to specifically targeting and downregulating the Minc03328 gene during nematode parasitism. Transgenic Minc03328-dsRNA lines that significantly downregulated Minc03328 gene expression during M. incognita parasitism were significantly less susceptible. The number of galls, egg masses, and [galls/egg masses] ratio were reduced in these transgenic lines by up to 85%, 90%, and 87%, respectively. Transgenic Minc03328-dsRNA lines showed the presence of fewer and smaller galls, indicating that parasitism was hindered. Overall, data herein strongly suggest that Minc03328 effector protein is important for M. incognita parasitism establishment. As well, the in planta Minc03328-dsRNA strategy demonstrated high biotechnological potential for developing crop species that could efficiently control RKN in the field.

Strategies to Enhance Periplasmic Recombinant Protein Production Yields in Escherichia coli

Main reasons to produce recombinant proteins in the periplasm of E. coli rather than in its cytoplasm are to -i- enable disulfide bond formation, -ii- facilitate protein isolation, -iii- control the nature of the N-terminus of the mature protein, and -iv- minimize exposure to cytoplasmic proteases. However, hampered protein targeting, translocation and folding as well as protein instability can all negatively affect periplasmic protein production yields. Strategies to enhance periplasmic protein production yields have focused on harmonizing secretory recombinant protein production rates with the capacity of the secretory apparatus by transcriptional and translational tuning, signal peptide selection and engineering, increasing the targeting, translocation and periplasmic folding capacity of the production host, preventing proteolysis, and, finally, the natural and engineered adaptation of the production host to periplasmic protein production. Here, we discuss these strategies using notable examples as a thread.

Phosphorylation of Extracellular Proteins in Acinetobacter baumannii in Sessile Mode of Growth

Acinetobacter baumannii is a problematic nosocomial pathogen owing to its increasing resistance to antibiotics and its great ability to survive in the hospital environment, which is linked to its capacity to form biofilms. Structural and functional investigations of post-translational modifications, such as phosphorylations, may lead to identification of candidates for therapeutic targets against this pathogen. Here, we present the first S/T/Y phosphosecretome of two A. baumannii strains, the reference strain ATCC 17978 and the virulent multi-drug resistant strain AB0057, cultured in two modes of growth (planktonic and biofilm) using TiO2 chromatography followed by high resolution mass spectrometry. In ATCC 17978, we detected a total of 137 (97 phosphoproteins) and 52 (33 phosphoproteins) phosphosites in biofilm and planktonic modes of growth, respectively. Similarly, in AB0057, 155 (119 phosphoproteins) and 102 (74 phosphoproteins) phosphosites in biofilm and planktonic modes of growth were identified, respectively. Both strains in the biofilm mode of growth showed a higher number of phosphosites and phosphoproteins compared to planktonic growth. Several phosphorylated sites are localized in key regions of proteins involved in either drug resistance (β-lactamases), adhesion to host tissues (pilins), or protein secretion (Hcp). Site-directed mutagenesis of the Hcp protein, essential for type VI secretion system-mediated interbacterial competition, showed that four of the modified residues are essential for type VI secretion system activity.

Automated machine learning optimizes and accelerates predictive modeling from COVID-19 high throughput datasets

COVID-19 outbreak brings intense pressure on healthcare systems, with an urgent demand for effective diagnostic, prognostic and therapeutic procedures. Here, we employed Automated Machine Learning (AutoML) to analyze three publicly available high throughput COVID-19 datasets, including proteomic, metabolomic and transcriptomic measurements. Pathway analysis of the selected features was also performed. Analysis of a combined proteomic and metabolomic dataset led to 10 equivalent signatures of two features each, with AUC 0.840 (CI 0.723-0.941) in discriminating severe from non-severe COVID-19 patients. A transcriptomic dataset led to two equivalent signatures of eight features each, with AUC 0.914 (CI 0.865-0.955) in identifying COVID-19 patients from those with a different acute respiratory illness. Another transcriptomic dataset led to two equivalent signatures of nine features each, with AUC 0.967 (CI 0.899-0.996) in identifying COVID-19 patients from virus-free individuals. Signature predictive performance remained high upon validation. Multiple new features emerged and pathway analysis revealed biological relevance by implication in Viral mRNA Translation, Interferon gamma signaling and Innate Immune System pathways. In conclusion, AutoML analysis led to multiple biosignatures of high predictive performance, with reduced features and large choice of alternative predictors. These favorable characteristics are eminent for development of cost-effective assays to contribute to better disease management.

Prediction of outcome in patients with non-small cell lung cancer treated with second line PD-1/PDL-1 inhibitors based on clinical parameters: Results from a prospective, single institution study

OBJECTIVE

We prospectively recorded clinical and laboratory parameters from patients with metastatic non-small cell lung cancer (NSCLC) treated with 2nd line PD-1/PD-L1 inhibitors in order to address their effect on treatment outcomes.

MATERIALS AND METHODS

Clinicopathological information (age, performance status, smoking, body mass index, histology, organs with metastases), use and duration of proton pump inhibitors, steroids and antibiotics (ATB) and laboratory values [neutrophil/lymphocyte ratio, LDH, albumin] were prospectively collected. Steroid administration was defined as the use of > 10 mg prednisone equivalent for ≥ 10 days. Prolonged ATB administration was defined as ATB ≥ 14 days 30 days before or within the first 3 months of treatment. JADBio, a machine learning pipeline was applied for further multivariate analysis.

RESULTS

Data from 66 pts with non-oncogenic driven metastatic NSCLC were analyzed; 15.2% experienced partial response (PR), 34.8% stable disease (SD) and 50% progressive disease (PD). Median overall survival (OS) was 6.77 months. ATB administration did not affect patient OS [HR = 1.35 (CI: 0.761-2.406, p = 0.304)], however, prolonged ATBs [HR = 2.95 (CI: 1.62-5.36, p = 0.0001)] and the presence of bone metastases [HR = 1.89 (CI: 1.02-3.51, p = 0.049)] independently predicted for shorter survival. Prolonged ATB administration, bone metastases, liver metastases and BMI < 25 kg/m2 were selected by JADbio as the important features that were associated with increased probability of developing disease progression as response to treatment. The resulting algorithm that was created was able to predict the probability of disease stabilization (PR or SD) in a single individual with an AUC = 0.806 [95% CI:0.714-0.889].

CONCLUSIONS

Our results demonstrate an adverse effect of prolonged ATBs on response and survival and underscore their importance along with the presence of bone metastases, liver metastases and low BMI in the individual prediction of outcomes in patients treated with immunotherapy.

Accurate Blood-Based Diagnostic Biosignatures for Alzheimer's Disease via Automated Machine Learning

Alzheimer's disease (AD) is the most common form of neurodegenerative dementia and its timely diagnosis remains a major challenge in biomarker discovery. In the present study, we analyzed publicly available high-throughput low-sample -omics datasets from studies in AD blood, by the AutoML technology Just Add Data Bio (JADBIO), to construct accurate predictive models for use as diagnostic biosignatures. Considering data from AD patients and age-sex matched cognitively healthy individuals, we produced three best performing diagnostic biosignatures specific for the presence of AD: A. A 506-feature transcriptomic dataset from 48 AD and 22 controls led to a miRNA-based biosignature via Support Vector Machines with three miRNA predictors (AUC 0.975 (0.906, 1.000)), B. A 38,327-feature transcriptomic dataset from 134 AD and 100 controls led to six mRNA-based statistically equivalent signatures via Classification Random Forests with 25 mRNA predictors (AUC 0.846 (0.778, 0.905)) and C. A 9483-feature proteomic dataset from 25 AD and 37 controls led to a protein-based biosignature via Ridge Logistic Regression with seven protein predictors (AUC 0.921 (0.849, 0.972)). These performance metrics were also validated through the JADBIO pipeline confirming stability. In conclusion, using the automated machine learning tool JADBIO, we produced accurate predictive biosignatures extrapolating available low sample -omics data. These results offer options for minimally invasive blood-based diagnostic tests for AD, awaiting clinical validation based on respective laboratory assays. They also highlight the value of AutoML in biomarker discovery.

Evolution of Escherichia coli Expression System in Producing Antibody Recombinant Fragments

Antibodies and antibody-derived molecules are continuously developed as both therapeutic agents and key reagents for advanced diagnostic investigations. Their application in these fields has indeed greatly expanded the demand of these molecules and the need for their production in high yield and purity. While full-length antibodies require mammalian expression systems due to the occurrence of functionally and structurally important glycosylations, most antibody fragments and antibody-like molecules are non-glycosylated and can be more conveniently prepared in E. coli-based expression platforms. We propose here an updated survey of the most effective and appropriate methods of preparation of antibody fragments that exploit E. coli as an expression background and review the pros and cons of the different platforms available today. Around 250 references accompany and complete the review together with some lists of the most important new antibody-like molecules that are on the market or are being developed as new biotherapeutics or diagnostic agents.

Applicability of an Automated Model and Parameter Selection in the Prediction of Screening-Level PTSD in Danish Soldiers Following Deployment: Development Study of Transferable Predictive Models Using Automated Machine Learning

BACKGROUND

Posttraumatic stress disorder (PTSD) is a relatively common consequence of deployment to war zones. Early postdeployment screening with the aim of identifying those at risk for PTSD in the years following deployment will help deliver interventions to those in need but have so far proved unsuccessful.

OBJECTIVE

This study aimed to test the applicability of automated model selection and the ability of automated machine learning prediction models to transfer across cohorts and predict screening-level PTSD 2.5 years and 6.5 years after deployment.

METHODS

Automated machine learning was applied to data routinely collected 6-8 months after return from deployment from 3 different cohorts of Danish soldiers deployed to Afghanistan in 2009 (cohort 1, N=287 or N=261 depending on the timing of the outcome assessment), 2010 (cohort 2, N=352), and 2013 (cohort 3, N=232).

RESULTS

Models transferred well between cohorts. For screening-level PTSD 2.5 and 6.5 years after deployment, random forest models provided the highest accuracy as measured by area under the receiver operating characteristic curve (AUC): 2.5 years, AUC=0.77, 95% CI 0.71-0.83; 6.5 years, AUC=0.78, 95% CI 0.73-0.83. Linear models performed equally well. Military rank, hyperarousal symptoms, and total level of PTSD symptoms were highly predictive.

CONCLUSIONS

Automated machine learning provided validated models that can be readily implemented in future deployment cohorts in the Danish Defense with the aim of targeting postdeployment support interventions to those at highest risk for developing PTSD, provided the cohorts are deployed on similar missions.

Understanding the mammalian TRAP complex function(s)

In eukaryotic cells, about one-third of the synthesized proteins are translocated into the endoplasmic reticulum; they are membrane or lumen resident proteins and proteins direct to the Golgi apparatus. The co-translational translocation takes place through the heterotrimeric protein-conducting channel Sec61 which is associated with the ribosome and many accessory components, such as the heterotetrameric translocon-associated protein (TRAP) complex. Recently, microscopic techniques, such as cryo-electron microscopy and cryo-electron tomography, have enabled the determination of the translocation machinery structure. However, at present, there is a lack of understanding regarding the roles of some of its components; indeed, the TRAP complex function during co-translational translocation needs to be established. In addition, TRAP may play a role during unfolded protein response, endoplasmic-reticulum-associated protein degradation and congenital disorder of glycosylation (ssr4 CDG). In this article, I describe the current understanding of the TRAP complex in the light of its possible function(s).

Trigger factor is a bona fide secretory pathway chaperone that interacts with SecB and the translocase

Bacterial secretory preproteins are translocated across the inner membrane post-translationally by the SecYEG-SecA translocase. Mature domain features and signal peptides maintain preproteins in kinetically trapped, largely soluble, folding intermediates. Some aggregation-prone preproteins require chaperones, like trigger factor (TF) and SecB, for solubility and/or targeting. TF antagonizes the contribution of SecB to secretion by an unknown molecular mechanism. We reconstituted this interaction in vitro and studied targeting and secretion of the model preprotein pro-OmpA. TF and SecB display distinct, unsuspected roles in secretion. Tightly associating TF:pro-OmpA targets the translocase at SecA, but TF prevents pro-OmpA secretion. In solution, SecB binds TF:pro-OmpA with high affinity. At the membrane, when bound to the SecA C-tail, SecB increases TF and TF:pro-OmpA affinities for the translocase and allows pro-OmpA to resume translocation. Our data reveal that TF, a main cytoplasmic folding pathway chaperone, is also a bona fide post-translational secretory chaperone that directly interacts with both SecB and the translocase to mediate regulated protein secretion. Thus, TF links the cytoplasmic folding and secretion chaperone networks.

Structural Basis of the Subcellular Topology Landscape of Escherichia coli

Cellular proteomes are distributed in multiple compartments: on DNA, ribosomes, on and inside membranes, or they become secreted. Structural properties that allow polypeptides to occupy subcellular niches, particularly to after crossing membranes, remain unclear. We compared intrinsic and extrinsic features in cytoplasmic and secreted polypeptides of the Escherichia coli K-12 proteome. Structural features between the cytoplasmome and secretome are sharply distinct, such that a signal peptide-agnostic machine learning tool distinguishes cytoplasmic from secreted proteins with 95.5% success. Cytoplasmic polypeptides are enriched in aliphatic, aromatic, charged and hydrophobic residues, unique folds and higher early folding propensities. Secretory polypeptides are enriched in polar/small amino acids, β folds, have higher backbone dynamics, higher disorder and contact order and are more often intrinsically disordered. These non-random distributions and experimental evidence imply that evolutionary pressure selected enhanced secretome flexibility, slow folding and looser structures, placing the secretome in a distinct protein class. These adaptations protect the secretome from premature folding during its cytoplasmic transit, optimize its lipid bilayer crossing and allowed it to acquire cell envelope specific chemistries. The latter may favor promiscuous multi-ligand binding, sensing of stress and cell envelope structure changes. In conclusion, enhanced flexibility, slow folding, looser structures and unique folds differentiate the secretome from the cytoplasmome. These findings have wide implications on the structural diversity and evolution of modern proteomes and the protein folding problem.

Enhancing Recombinant Protein Yields in the E. coli Periplasm by Combining Signal Peptide and Production Rate Screening

Proteins that contain disulfide bonds mainly mature in the oxidative environment of the eukaryotic endoplasmic reticulum or the periplasm of Gram-negative bacteria. In E. coli, disulfide bond containing recombinant proteins are often targeted to the periplasm by an N-terminal signal peptide that is removed once it passes through the Sec-translocon in the cytoplasmic membrane. Despite their conserved targeting function, signal peptides can impact recombinant protein production yields in the periplasm, as can the production rate. Here, we present a combined screen involving different signal peptides and varying production rates that enabled the identification of more optimal conditions for periplasmic production of recombinant proteins with disulfide bonds. The data was generated from two targets, a single chain antibody fragment (BL1) and human growth hormone (hGH), with four different signal peptides and a titratable rhamnose promoter-based system that enables the tuning of protein production rates. Across the screen conditions, the yields for both targets significantly varied, and the optimal signal peptide and rhamnose concentration differed for each protein. Under the optimal conditions, the periplasmic BL1 and hGH were properly folded and active. Our study underpins the importance of combinatorial screening approaches for addressing the requirements associated with the production of a recombinant protein in the periplasm.

Regulating Protein Function by Delayed Folding

In this issue of Structure, Tsirigotaki et al. (2018) use bioinformatics and biophysical tools to demonstrate that many secreted proteins form long-lived, loosely packed folding intermediates. This delayed folding correlates with elevated disorder and reduced hydrophobicity compared to structured cytosolic proteins and is often stabilized by signal peptides by yet to be determined mechanisms.

Circulating cell-free DNA in breast cancer: size profiling, levels, and methylation patterns lead to prognostic and predictive classifiers

Blood circulating cell-free DNA (ccfDNA) is a suggested biosource of valuable clinical information for cancer, meeting the need for a minimally-invasive advancement in the route of precision medicine. In this paper, we evaluated the prognostic and predictive potential of ccfDNA parameters in early and advanced breast cancer. Groups consisted of 150 and 16 breast cancer patients under adjuvant and neoadjuvant therapy respectively, 34 patients with metastatic disease and 35 healthy volunteers. Direct quantification of ccfDNA in plasma revealed elevated concentrations correlated to the incidence of death, shorter PFS, and non-response to pharmacotherapy in the metastatic but not in the other groups. The methylation status of a panel of cancer-related genes chosen based on previous expression and epigenetic data (KLK10, SOX17, WNT5A, MSH2, GATA3) was assessed by quantitative methylation-specific PCR. All but the GATA3 gene was more frequently methylated in all the patient groups than in healthy individuals (all p < 0.05). The methylation of WNT5A was statistically significantly correlated to greater tumor size and poor prognosis characteristics and in advanced stage disease with shorter OS. In the metastatic group, also SOX17 methylation was significantly correlated to the incidence of death, shorter PFS, and OS. KLK10 methylation was significantly correlated to unfavorable clinicopathological characteristics and relapse, whereas in the adjuvant group to shorter DFI. Methylation of at least 3 or 4 genes was significantly correlated to shorter OS and no pharmacotherapy response, respectively. Classification analysis by a fully automated, machine learning software produced a single-parametric linear model using ccfDNA plasma concentration values, with great discriminating power to predict response to chemotherapy (AUC 0.803, 95% CI [0.606, 1.000]) in the metastatic group. Two more multi-parametric signatures were produced for the metastatic group, predicting survival and disease outcome. Finally, a multiple logistic regression model was constructed, discriminating between patient groups and healthy individuals. Overall, ccfDNA emerged as a highly potent predictive classifier in metastatic breast cancer. Upon prospective clinical evaluation, all the signatures produced could aid accurate prognosis.

Toward Automatic Risk Assessment to Support Suicide Prevention

Background: Suicide has been considered an important public health issue for years and is one of the main causes of death worldwide. Despite prevention strategies being applied, the rate of suicide has not changed substantially over the past decades. Suicide risk has proven extremely difficult to assess for medical specialists, and traditional methodologies deployed have been ineffective. Advances in machine learning make it possible to attempt to predict suicide with the analysis of relevant data aiming to inform clinical practice. Aims: We aimed to (a) test our artificial intelligence based, referral-centric methodology in the context of the National Health Service (NHS), (b) determine whether statistically relevant results can be derived from data related to previous suicides, and (c) develop ideas for various exploitation strategies. Method: The analysis used data of patients who died by suicide in the period 2013-2016 including both structured data and free-text medical notes, necessitating the deployment of state-of-the-art machine learning and text mining methods. Limitations: Sample size is a limiting factor for this study, along with the absence of non-suicide cases. Specific analytical solutions were adopted for addressing both issues. Results and Conclusion: The results of this pilot study indicate that machine learning shows promise for predicting within a specified period which people are most at risk of taking their own life at the time of referral to a mental health service.

Bootstrapping the out-of-sample predictions for efficient and accurate cross-validation

Cross-Validation (CV), and out-of-sample performance-estimation protocols in general, are often employed both for (a) selecting the optimal combination of algorithms and values of hyper-parameters (called a configuration) for producing the final predictive model, and (b) estimating the predictive performance of the final model. However, the cross-validated performance of the best configuration is optimistically biased. We present an efficient bootstrap method that corrects for the bias, called Bootstrap Bias Corrected CV (BBC-CV). BBC-CV's main idea is to bootstrap the whole process of selecting the best-performing configuration on the out-of-sample predictions of each configuration, without additional training of models. In comparison to the alternatives, namely the nested cross-validation (Varma and Simon in BMC Bioinform 7(1):91, 2006) and a method by Tibshirani and Tibshirani (Ann Appl Stat 822-829, 2009), BBC-CV is computationally more efficient, has smaller variance and bias, and is applicable to any metric of performance (accuracy, AUC, concordance index, mean squared error). Subsequently, we employ again the idea of bootstrapping the out-of-sample predictions to speed up the CV process. Specifically, using a bootstrap-based statistical criterion we stop training of models on new folds of inferior (with high probability) configurations. We name the method Bootstrap Bias Corrected with Dropping CV (BBCD-CV) that is both efficient and provides accurate performance estimates.

Long-Lived Folding Intermediates Predominate the Targeting-Competent Secretome

Secretory preproteins carry signal peptides fused amino-terminally to mature domains. They are post-translationally targeted to cross the plasma membrane in non-folded states with the help of translocases, and fold only at their final destinations. The mechanism of this process of postponed folding is unknown, but is generally attributed to signal peptides and chaperones. We herein demonstrate that, during targeting, most mature domains maintain loosely packed folding intermediates. These largely soluble states are signal peptide independent and essential for translocase recognition. These intermediates are promoted by mature domain features: residue composition, elevated disorder, and reduced hydrophobicity. Consequently, a mature domain folds slower than its cytoplasmic structural homolog. Some mature domains could not evolve stable, loose intermediates, and hence depend on signal peptides for slow folding to the detriment of solubility. These unique features of secretory proteins impact our understanding of protein trafficking, folding, and aggregation, and thus place them in a distinct class.