Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now

×

Discipline
Biological
Keywords
Chloroplast
Pre-Sequence
Targeting
Thalassiosira Pseudonana
Phaeodactylum Tricornutum
Observation Type
Follow-up
Nature
Resources / Big Data
Submitted
Apr 9th, 2019
Published
Jun 18th, 2020
  • Abstract

    Plastid targeted proteins of diatoms and related algae can be predicted with high sensitivity and specificity using the ASAFind method published in 2015. ASAFind predictions rely on SignalP predictions of endoplasmic reticulum (ER) targeting signal peptides. Recently (in 2019), a new version of SignalP was released, SignalP 5.0. We tested the ability of SignalP 5.0 to recognize signal peptides of nucleus-encoded, plastid-targeted diatom pre-proteins, and to identify the signal peptide cleavage site. The results were compared to manual predictions of the characteristic cleavage site motif, and to previous versions of SignalP. SignalP 5.0 is less sensitive than the previous versions of SignalP in this specific task, and also in the detection of signal peptides of non-plastid proteins in diatoms. However, in combination with ASAFind, the resulting prediction performance for plastid proteins is high. In addition, we tested the multi-location prediction tool TargetP for its suitability to provide signal peptide information to ASAFind. The newest version, TargetP 2.0, had the highest prediction performances for diatom signal peptides and mitochondrial transit peptides compared to other versions of SignalP and TargetP, thus it provides a good basis for ASAFind predictions.

  • Figure
  • Introduction

    Diatoms are important primary producers in many aquatic habitats and are increasingly important for biotechnological applications. Genome and transcriptome sequencing has revealed unusual aspects of diatom metabolism, many of which are related to the intracellular location of metabolic pathways in the diatom cell. Predicting intracellular locations of proteins from sequence data is therefore an important step in diatom genome annotation.

    Plastid proteins in diatoms are either plastid-encoded or are nucleus-encoded and need to be targeted to the complex diatom plastids, across four envelope membranes. Nucleus encoded plastid proteins of diatoms are recognized by characteristic bipartite targeting signals, consisting of a signal peptide domain, a transit peptide domain, a conserved motif at the signal peptide cleavage site, and another conserved motif at the transit peptide cleavage site.

    Plastid proteomes of two species of diatoms have been predicted with the dedicated prediction program ASAFind, and new sequence data can be analyzed with the ASAFind web service or a downloadable software. ASAFind has also been used to predict plastid proteins of the cryptophyte Guillardia theta, and other organisms with complex plastids of red algal origin. ASAFind builds upon the prediction of ER signal peptides by SignalP, a software that was released in different versions, which employ different prediction algorithms and therefore also differ in their predictions and performance.

  • Objective

    Since the release of the ASAFind prediction method in 2015, new versions of SignalP and TargetP have been released. The objective of this study is to find out whether these updated software tools improve the prediction of diatom plastid targeting pre-sequences with ASAFind, in comparison to the versions of SignalP with which ASAFind was originally designed to work.

  • Results & Discussion

    The ASAFind plastid protein prediction is based on the prediction of signal peptides by the prediction software SignalP. Predictions of signal peptides, as well as positions of predicted cleavage sites in diatom plastid targeting pre-sequences, differ between the SignalP versions available in 2007. Because ASAFind performs a sliding window motif search surrounding the position of the SignalP predicted signal peptide cleavage site, the choice of the most suitable version of SignalP is crucial for the performance of ASAFind. We tested the training set of 166 putative plastid-targeted protein sequences used to build the scoring matrix for the ASAFind method with the newly released SignalP 5.0 and with two versions of TargetP, with which ASAFind has not been tested. SignalP 5.0 predicted fewer signal peptides than most previous versions of SignalP (149 of the 166 signal peptides). TargetP 1.1 predicted only 142 signal peptides while TargetP 2.0 predicted 162 signal peptides (see Fig. 1A and corresponding raw data).

    In addition to the prediction of a signal peptide, SignalP and TargetP also predict the position of the signal peptide cleavage site. ASAFind uses the predicted position of the cleavage site to determine the range of the sequence windows in which the sequence is scored according to the conserved motif. This is in part because there is considerable variation in the predicted positions of the signal peptide cleavage sites between SignalP 3.0 NN, HMM, and SignalP 4.1, the versions of SignalP that were available in 2015. SignalP 5.0, TargetP 1.1, and TargetP 2.0 also differ in the prediction of the cleavage site position between sequences (see Fig. 1A and corresponding raw data). Notably, while there are some sequences where the predicted signal peptide cleavage site is different from the manually predicted one across several versions of the prediction programs, there are many other examples of individual sequences for which different versions of the prediction programs give different results.

    Due to the sliding window approach of ASAFind, the exact position of the SignalP predicted cleavage site is less important for the prediction result than the actual score of the highest-scoring sequence window. We therefore also tested the performance of ASAFind in combination with SignalP 5.0, TargetP 1.1, and TargetP 2.0 using the reference set of 132 experimentally localized proteins from various intracellular locations in Phaeodactylum tricornutum (Table S3 of). Sensitivities, specificities, and MCCs are compiled in figure 1B (raw data also attached). For plastid proteins, the highest MCC is obtained by combining ASAFind with the SignalP 4.1 "sensitive" setting (Fig. 1B). For overall signal peptide predictions (of plastid as well as non-plastid proteins), and also for mitochondrial transit peptide predictions, TargetP 2.0 has the best performance by MCC (Fig. 1B). Because TargetP 2.0 also performs well specifically for plastid proteins in combination with ASAFind (Fig. 1B), TargetP 2.0 is a valuable tool for protein targeting analyses in diatoms.

  • Conclusions

    SignalP 5.0 is a highly specific prediction program for signal peptides and differentiates between three types of signal peptides in Bacteria and Archaea. While its overall performance in the organisms for which it was developed is higher than that of its predecessors, our results show that it is at the same time, less sensitive in the detection of signal peptides in proteins from diatoms. Nevertheless, it gives good results for diatom plastid protein prediction in conjunction with ASAFind. TargetP 2.0 has the highest Matthews correlation coefficients for the prediction of signal peptides and mitochondrial transit peptides in diatoms and thus is also suitable as a basis for ASAFind plastid protein predictions.

  • Limitations

    The 132 protein reference set of was originally compiled with the goal to evaluate the performance of plastid protein predictions (55 positives, 77 negatives, which contained all experimental protein localizations in Phaeodactylum tricornutum published at the time). It is highly asymmetric with respect to other sequence categories. For mitochondria, it only contains 10 positive sequences (versus 122 negatives), and for signal peptides, it only contains 17 negative sequences (versus 115 positive sequences, including the 55 plastid-targeted proteins). These ratios can lead to over-optimistic performance measure in the case of sensitivity and specificity. We therefore also included MCCs, which account for asymmetries in the reference set. The MCCs are low for the signal peptide and mitochondrial transit peptide predictions tested with this dataset, in comparison to the MCCs for plastid protein predictions (Fig. 1B).

  • Alt. Explanations

    The prediction performances of SignalP and TargetP for signal peptides and mitochondrial transit peptides of diatoms might actually be higher than what the low MCCs (Fig. 1B) suggest, if they could be tested with more symmetric reference datasets (see "Limitations").

  • Conjectures

    In the future, larger reference sets will certainly help to improve the characteristics of the prediction methods. Additional studies with experimental protein localizations in diatoms have been published since 2015 (e.g.), and should be included in future reference sets. Here, we used the reference set from, in order to keep the results comparable to the previously published study.

  • Methods

    The sequences used to calculate the ASAFind scoring matrices (training set, Table S1 from) and the reference protein set (Table S3 from) were downloaded from the journal homepage. For these sequences, prediction results of SignalP 3.0 NN, SignalP 3.0 HMM, SignalP 4.1 (with default and "sensitive" settings) and SignalP 5.0 (with the "Eukarya" organism group selected) were obtained from the SignalP web server () (if not already contained in Table S1 from). TargetP results were obtained from the TargetP web servers (http://www.cbs.dtu.dk/services/TargetP/) for TargetP 1.1 (with the default "no cutoffs; winner-takes-all" option, and the "non-plant" organism group selected) and TargetP 2.0 ("non-plant" organism group selected). In both TargetP variants, "non-plant" should be used as the organism group, because otherwise, TargetP will also search for chloroplast transit peptides known from primary plastids of higher plants and green algae, a type of targeting pre-sequence that does not exist in diatoms. The results were re-formatted to ASAFind input requirements with spreadsheet software and a text editor. For SignalP 5.0 and TargetP 2.0, the content of the "Prediction" and "CS Position" columns of the output table was used to start ASAFind scoring of the sequences with predicted signal peptides; for TargetP 1.1, the content of the "Loc" and "TPlen" columns was used (with "TPlen" corrected by +1 to obtain the cleavage site position). The collected results were analyzed using spreadsheet software and custom Perl scripts. The performance statistics of HECTAR on the reference set was reproduced from. For the calculation of the prediction performance of signal peptides, all sequences from the reference set that enter the secretory pathway were counted as "positive", all others were counted as "negative"; for the calculation of the prediction performance of plastid proteins, all experimentally plastid localized proteins from the reference set were counted as "positive", all others were counted as "negative"; for the calculation of the prediction performance of mitochondrial proteins, all experimentally mitochondria localized proteins from the reference set were counted as "positive", all others were counted as "negative" (see raw data file for the exact assignments). Sensitivities, specificities, and Matthews correlation coefficients (MCC) were calculated according to the formulas given in.

  • Funding statement

    Our research is supported by the Institute of Parasitology of the Czech Academy of Sciences, by the Grant Agency of the Czech Republic (18-13458S), by the ERDF/ESF Centre for research of pathogenicity and virulence of parasites (No.CZ.02.1.01/0.0/0.0/16_019/0000759) and by the US National Science Foundation (G.R and C.M).

  • Ethics statement

    Not Applicable.

  • References
  • 1
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    2
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    3
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    4
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    5
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    Matters10/20

    Comparison of different versions of SignalP and TargetP for diatom plastid protein predictions with ASAFind

    Affiliation listing not available.
    Repository

  • Ansgar Gruber, Gabrielle Rocap, Peter G. Kroth, E. Virginia Armbrust, Thomas Mock
    Plastid proteome prediction for diatoms and other algae with secondary plastids of the red lineage
    The Plant Journal, 81/2015, pages 519-528 DOI: 10.1111/tpj.12734chrome_reader_mode
  • Abstractlink

    Plastid targeted proteins of diatoms and related algae can be predicted with high sensitivity and specificity using the ASAFind method published in 2015. ASAFind predictions rely on SignalP predictions of endoplasmic reticulum (ER) targeting signal peptides. Recently (in 2019), a new version of SignalP was released, SignalP 5.0. We tested the ability of SignalP 5.0 to recognize signal peptides of nucleus-encoded, plastid-targeted diatom pre-proteins, and to identify the signal peptide cleavage site. The results were compared to manual predictions of the characteristic cleavage site motif, and to previous versions of SignalP. SignalP 5.0 is less sensitive than the previous versions of SignalP in this specific task, and also in the detection of signal peptides of non-plastid proteins in diatoms. However, in combination with ASAFind, the resulting prediction performance for plastid proteins is high. In addition, we tested the multi-location prediction tool TargetP for its suitability to provide signal peptide information to ASAFind. The newest version, TargetP 2.0, had the highest prediction performances for diatom signal peptides and mitochondrial transit peptides compared to other versions of SignalP and TargetP, thus it provides a good basis for ASAFind predictions.

    Figurelink

    Figure 1A.

    Comparison of results from prediction servers with manually predicted signal peptide cleavage site positions in the training set of 83 orthologous pairs of putative plastid-targeted protein sequences from Thalassiosira pseudonana and Phaeodactylum tricornutum used to generate the ASAFind scoring matrix (found in Table S1 of[1], raw data from the prediction servers are attached). SignalP 3.0 uses either a neural network (NN) or a hidden Markov (HMM) model. SignalP 4.1 was tested with the default (def) and with the sensitive (sen) options, SignalP 5.0 employs a deep neural network, see methods for references. Sequences are plotted in the same order as they appear in the raw data files. The numbers below each prediction server indicate the number of cleavage sites that matched the manual prediction/ and the total number of signal peptides predicted.

    Figure 1B.

    Sensitivities, specificities and Matthews correlation coefficients (MCC) of the different variations of SignalP and TargetP for the prediction of signal peptides, mitochondrial transit peptides and plastid proteins (in conjunction with ASAFind), determined with a set of 132 experimentally localized reference proteins from multiple intracellular destinations in Phaeodactylum tricornutum (found in Table S3 of[1]). Results are sorted from highest to lowest MCC. The performance of HECTAR[2] is also listed for comparison.

    Introductionlink

    Diatoms are important primary producers in many aquatic habitats[3] and are increasingly important for biotechnological applications[4][5]. Genome and transcriptome sequencing has revealed unusual aspects of diatom metabolism, many of which are related to the intracellular location of metabolic pathways in the diatom cell[6]. Predicting intracellular locations of proteins from sequence data is therefore an important step in diatom genome annotation[7][8][9].

    Plastid proteins in diatoms are either plastid-encoded or are nucleus-encoded and need to be targeted to the complex diatom plastids, across four envelope membranes[10][11]. Nucleus encoded plastid proteins of diatoms are recognized by characteristic bipartite targeting signals, consisting of a signal peptide domain, a transit peptide domain, a conserved motif at the signal peptide cleavage site, and another conserved motif at the transit peptide cleavage site[12][13][14][15][10][16][17][18][19][20][21].

    Plastid proteomes of two species of diatoms have been predicted with the dedicated prediction program ASAFind, and new sequence data can be analyzed with the ASAFind web service or a downloadable software[1]. ASAFind has also been used to predict plastid proteins of the cryptophyte Guillardia theta[22], and other organisms with complex plastids of red algal origin[23][24][25]. ASAFind builds upon the prediction of ER signal peptides by SignalP, a software that was released in different versions, which employ different prediction algorithms and therefore also differ in their predictions and performance[26][27][28].

    Objectivelink

    Since the release of the ASAFind prediction method in 2015[1], new versions of SignalP[29] and TargetP[30] have been released. The objective of this study is to find out whether these updated software tools improve the prediction of diatom plastid targeting pre-sequences with ASAFind, in comparison to the versions of SignalP with which ASAFind was originally designed to work.

    Results & Discussionlink

    The ASAFind plastid protein prediction[1] is based on the prediction of signal peptides by the prediction software SignalP[26][28]. Predictions of signal peptides, as well as positions of predicted cleavage sites in diatom plastid targeting pre-sequences, differ between the SignalP versions available in 2007[18]. Because ASAFind performs a sliding window motif search surrounding the position of the SignalP predicted signal peptide cleavage site, the choice of the most suitable version of SignalP is crucial for the performance of ASAFind[1]. We tested the training set of 166 putative plastid-targeted protein sequences used to build the scoring matrix for the ASAFind method with the newly released SignalP 5.0[29] and with two versions of TargetP[31][32][30], with which ASAFind has not been tested. SignalP 5.0 predicted fewer signal peptides than most previous versions of SignalP (149 of the 166 signal peptides). TargetP 1.1 predicted only 142 signal peptides while TargetP 2.0 predicted 162 signal peptides (see Fig. 1A and corresponding raw data).

    In addition to the prediction of a signal peptide, SignalP and TargetP also predict the position of the signal peptide cleavage site[26]. ASAFind uses the predicted position of the cleavage site to determine the range of the sequence windows in which the sequence is scored according to the conserved motif[1]. This is in part because there is considerable variation in the predicted positions of the signal peptide cleavage sites between SignalP 3.0 NN, HMM, and SignalP 4.1, the versions of SignalP that were available in 2015[1]. SignalP 5.0, TargetP 1.1, and TargetP 2.0 also differ in the prediction of the cleavage site position between sequences (see Fig. 1A and corresponding raw data). Notably, while there are some sequences where the predicted signal peptide cleavage site is different from the manually predicted one across several versions of the prediction programs, there are many other examples of individual sequences for which different versions of the prediction programs give different results.

    Due to the sliding window approach of ASAFind, the exact position of the SignalP predicted cleavage site is less important for the prediction result than the actual score of the highest-scoring sequence window[1]. We therefore also tested the performance of ASAFind in combination with SignalP 5.0, TargetP 1.1, and TargetP 2.0 using the reference set of 132 experimentally localized proteins from various intracellular locations in Phaeodactylum tricornutum (Table S3 of[1]). Sensitivities, specificities, and MCCs are compiled in figure 1B (raw data also attached). For plastid proteins, the highest MCC is obtained by combining ASAFind with the SignalP 4.1 "sensitive" setting (Fig. 1B). For overall signal peptide predictions (of plastid as well as non-plastid proteins), and also for mitochondrial transit peptide predictions, TargetP 2.0 has the best performance by MCC (Fig. 1B). Because TargetP 2.0 also performs well specifically for plastid proteins in combination with ASAFind (Fig. 1B), TargetP 2.0 is a valuable tool for protein targeting analyses in diatoms.

    Conclusionslink

    SignalP 5.0 is a highly specific prediction program for signal peptides and differentiates between three types of signal peptides in Bacteria and Archaea. While its overall performance in the organisms for which it was developed is higher than that of its predecessors[29], our results show that it is at the same time, less sensitive in the detection of signal peptides in proteins from diatoms. Nevertheless, it gives good results for diatom plastid protein prediction in conjunction with ASAFind. TargetP 2.0[30] has the highest Matthews correlation coefficients for the prediction of signal peptides and mitochondrial transit peptides in diatoms and thus is also suitable as a basis for ASAFind plastid protein predictions.

    Limitationslink

    The 132 protein reference set of[1] was originally compiled with the goal to evaluate the performance of plastid protein predictions (55 positives, 77 negatives, which contained all experimental protein localizations in Phaeodactylum tricornutum published at the time). It is highly asymmetric with respect to other sequence categories. For mitochondria, it only contains 10 positive sequences (versus 122 negatives), and for signal peptides, it only contains 17 negative sequences (versus 115 positive sequences, including the 55 plastid-targeted proteins). These ratios can lead to over-optimistic performance measure in the case of sensitivity and specificity. We therefore also included MCCs, which account for asymmetries in the reference set. The MCCs are low for the signal peptide and mitochondrial transit peptide predictions tested with this dataset, in comparison to the MCCs for plastid protein predictions (Fig. 1B).

    Alternative Explanationslink

    The prediction performances of SignalP and TargetP for signal peptides and mitochondrial transit peptides of diatoms might actually be higher than what the low MCCs (Fig. 1B) suggest, if they could be tested with more symmetric reference datasets (see "Limitations").

    Conjectureslink

    In the future, larger reference sets will certainly help to improve the characteristics of the prediction methods. Additional studies with experimental protein localizations in diatoms have been published since 2015 (e.g.[33][34][35][36][37][38]), and should be included in future reference sets. Here, we used the reference set from[1], in order to keep the results comparable to the previously published study.

    Methodslink

    The sequences used to calculate the ASAFind scoring matrices (training set, Table S1 from[1]) and the reference protein set (Table S3 from[1]) were downloaded from the journal homepage. For these sequences, prediction results of SignalP 3.0 NN[39], SignalP 3.0 HMM[40], SignalP 4.1[41][27] (with default and "sensitive" settings) and SignalP 5.0[29] (with the "Eukarya" organism group selected) were obtained from the SignalP web server (http://www.cbs.dtu.dk/services/SignalP/) (if not already contained in Table S1 from[1]). TargetP results were obtained from the TargetP web servers (http://www.cbs.dtu.dk/services/TargetP/) for TargetP 1.1[31][32] (with the default "no cutoffs; winner-takes-all" option, and the "non-plant" organism group selected) and TargetP 2.0[30] ("non-plant" organism group selected). In both TargetP variants, "non-plant" should be used as the organism group, because otherwise, TargetP will also search for chloroplast transit peptides known from primary plastids of higher plants and green algae, a type of targeting pre-sequence that does not exist in diatoms. The results were re-formatted to ASAFind input requirements with spreadsheet software and a text editor. For SignalP 5.0 and TargetP 2.0, the content of the "Prediction" and "CS Position" columns of the output table was used to start ASAFind scoring of the sequences with predicted signal peptides; for TargetP 1.1, the content of the "Loc" and "TPlen" columns was used (with "TPlen" corrected by +1 to obtain the cleavage site position). The collected results were analyzed using spreadsheet software and custom Perl scripts. The performance statistics of HECTAR[2] on the reference set was reproduced from[1]. For the calculation of the prediction performance of signal peptides, all sequences from the reference set that enter the secretory pathway were counted as "positive", all others were counted as "negative"; for the calculation of the prediction performance of plastid proteins, all experimentally plastid localized proteins from the reference set were counted as "positive", all others were counted as "negative"; for the calculation of the prediction performance of mitochondrial proteins, all experimentally mitochondria localized proteins from the reference set were counted as "positive", all others were counted as "negative" (see raw data file for the exact assignments). Sensitivities, specificities, and Matthews correlation coefficients (MCC) were calculated according to the formulas given in[42].

    Funding Statementlink

    Our research is supported by the Institute of Parasitology of the Czech Academy of Sciences, by the Grant Agency of the Czech Republic (18-13458S), by the ERDF/ESF Centre for research of pathogenicity and virulence of parasites (No.CZ.02.1.01/0.0/0.0/16_019/0000759) and by the US National Science Foundation (G.R and C.M).

    Conflict of interestlink

    The authors declare no conflicts of interest.

    Ethics Statementlink

    Not Applicable.

    No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.

    Referenceslink
    1. Ansgar Gruber, Gabrielle Rocap, Peter G. Kroth, E. Virginia Armbrust, Thomas Mock
      Plastid proteome prediction for diatoms and other algae with secondary plastids of the red lineage
      The Plant Journal, 81/2015, pages 519-528 DOI: 10.1111/tpj.12734chrome_reader_mode
    2. Bernhard Gschloessl, Yann Guermeur, J Mark Cock
      HECTAR: A method to predict subcellular targeting in heterokonts
      BMC Bioinformatics, 9/2008, page 393 DOI: 10.1186/1471-2105-9-393chrome_reader_mode
    3. E. Virginia Armbrust
      The life of diatoms in the world's oceans
      Nature, 459/2009, pages 185-192 DOI: 10.1038/nature08057chrome_reader_mode
    4. Peter G. Kroth, Atle M. Bones, Fayza Daboussi, Maria I. Ferrante, Marianne Jaubert, Misha Kolot, Marianne Nymark, Carolina Río Bártulos, Andrés Ritter, Monia T. Russo, Manuel Serif, Per Winge, Angela Falciatore
      Genome editing in diatoms: achievements and goals
      Plant Cell Reports, 37/2018, pages 1401-1408 DOI: 10.1007/s00299-018-2334-1chrome_reader_mode
    5. Angela Falciatore, Marianne Jaubert, Jean-Pierre Bouly, Benjamin Bailleul, Thomas Mock
      Diatom Molecular Research Comes of Age: Model Species for Studying Phytoplankton Biology and Diversity
      The Plant Cell, 32/2020, pages 547-572 DOI: 10.1105/tpc.19.00158chrome_reader_mode
    6. Benoît Schoefs, Hanhua Hu, Peter G. Kroth
      The peculiar carbon metabolism in diatoms
      Philosophical Transactions of the Royal Society B: Biological Sciences, 372/2017, page 20160405 DOI: 10.1098/rstb.2016.0405chrome_reader_mode
    7. Ansgar Gruber, Peter G. Kroth
      Deducing Intracellular Distributions of Metabolic Pathways from Genomic Data
      Methods in Molecular Biology (Methods and Protocols): Plant Metabolism, 1083/2013, pages 187-211 DOI: 10.1007/978-1-62703-661-0_12chrome_reader_mode
    8. Ansgar Gruber, Peter G. Kroth
      Intracellular metabolic pathway distribution in diatoms and tools for genome-enabled experimental diatom research
      Philosophical Transactions of the Royal Society B: Biological Sciences, 372/2017, page 20160402 DOI: 10.1098/rstb.2016.0402chrome_reader_mode
    9. Daniel Moog
      In Silico Tools for the Prediction of Protein Import into Secondary Plastids
      Methods in Molecular Biology: Plastids, 1829/2018, pages 381-394 DOI: 10.1007/978-1-4939-8654-5_25chrome_reader_mode
    10. Peter G. Kroth
      Protein transport into secondary plastids and the evolution of primary and secondary plastids
      International Review of Cytology, 221/2002, pages 191-255 DOI: 10.1016/s0074-7696(02)21013-xchrome_reader_mode
    11. Sven B. Gould, Ross F. Waller, Geoffrey I. McFadden
      Plastid Evolution
      Annual Review of Plant Biology, 59/2008, pages 491-517 DOI: 10.1146/annurev.arplant.59.032607.092915chrome_reader_mode
    12. Devaki Bhaya, Arthur Grossman
      Targeting proteins to diatom plastids involves transport through an endoplasmic reticulum
      Molecular and General Genetics MGG, 229/1991, pages 400-404 DOI: 10.1007/bf00267462chrome_reader_mode
    13. Peter G. Pancic, Heinrich Strotmann
      Structure of the nuclear encoded γ subunit of CF0CF1 of the diatom Odontella sinensis including its presequence
      FEBS Letters, 320/1993, pages 61-66 DOI: 10.1016/0014-5793(93)81658-mchrome_reader_mode
    14. Markus Lang, Kirk E. Apt, Peter G. Kroth
      Protein Transport into “Complex” Diatom Plastids Utilizes Two Different Targeting Signals
      Journal of Biological Chemistry, 273/1998, pages 30973-30978 DOI: 10.1074/jbc.273.47.30973chrome_reader_mode
    15. Kirk E. Apt, Lioudmila Zaslavkaia, J. Casey Lippmeier, Markus Lang, Oliver Kilian, Rick Wetherbee, Arthur R. Grossman, Peter G. Kroth
      In vivo characterization of diatom multipartite plastid targeting signals
      Journal of Cell Science, 115/2002, pages 4061-4069 DOI: 10.1242/jcs.00092chrome_reader_mode
    16. E. Virginia Armbrust, John A. Berges, Chris Bowler, Beverley R. Green, Diego Martinez, Nicholas H. Putnam, Shiguo Zhou, Andrew E. Allen, Kirk E. Apt, Michael Bechner, Mark A. Brzezinski, Balbir K. Chaal, Anthony Chiovitti, Aubrey K. Davis, Mark S. Demarest, J. Chris Detter, Tijana Glavina, David Goodstein, Masood Z. Hadi, Uffe Hellsten, Mark Hildebrand, Bethany D. Jenkins, Jerzy Jurka, Vladimir V. Kapitonov, Nils Kröger, Winnie W. Y. Lau, Todd W. Lane, Frank W. Larimer, J. Casey Lippmeier, Susan Lucas, Mónica Medina, Anton Montsant, Miroslav Obornik, Micaela Schnitzler Parker, Brian Palenik, Gregory J. Pazour, Paul M. Richardson, Tatiana A. Rynearson, Mak A. Saito, David C. Schwartz, Kimberlee Thamatrakoln, Klaus Valentin, Assaf Vardi, Frances P. Wilkerson, Daniel S. Rokhsar
      The Genome of the Diatom Thalassiosira Pseudonana: Ecology, Evolution, and Metabolism
      Science, 306/2004, pages 79-86 DOI: 10.1126/science.1101156chrome_reader_mode
    17. Oliver Kilian, Peter G. Kroth
      Identification and characterization of a new conserved motif within the presequence of proteins targeted into complex diatom plastids
      The Plant Journal, 41/2004, pages 175-183 DOI: 10.1111/j.1365-313x.2004.02294.xchrome_reader_mode
    18. Ansgar Gruber, Sascha Vugrinec, Franziska Hempel, Sven B. Gould, Uwe-G. Maier, Peter G. Kroth
      Protein targeting into complex diatom plastids: functional characterisation of a specific targeting motif
      Plant Molecular Biology, 64/2007, pages 519-530 DOI: 10.1007/s11103-007-9171-xchrome_reader_mode
    19. Nicola J. Patron, Ross F. Waller
      Transit peptide diversity and divergence: A global analysis of plastid targeting signals
      Bioessays, 29/2007, pages 1048-1058 DOI: 10.1002/bies.20638chrome_reader_mode
    20. Gregor Felsner, Maik S Sommer, Uwe G Maier
      The physical and functional borders of transit peptide-like sequences in secondary endosymbionts
      BMC Plant Biology, 10/2010, page 223 DOI: 10.1186/1471-2229-10-223chrome_reader_mode
    21. Pitter F. Huesgen, Meriem Alami, Philipp F. Lange, Leonard J. Foster, Wolfgang P. Schröder, Christopher M. Overall, Beverley R. Green
      Proteomic Amino-Termini Profiling Reveals Targeting Information for Protein Import into Complex Plastids
    22. Bruce A. Curtis, Goro Tanifuji, Fabien Burki, Ansgar Gruber, Manuel Irimia, Shinichiro Maruyama, Maria C. Arias, Steven G. Ball, Gillian H. Gile, Yoshihisa Hirakawa, Julia F. Hopkins, Alan Kuo, Stefan A. Rensing, Jeremy Schmutz, Aikaterini Symeonidi, Marek Elias, Robert J. M. Eveleigh, Emily K. Herman, Mary J. Klute, Takuro Nakayama, Miroslav Oborník, Adrian Reyes-Prieto, E. Virginia Armbrust, Stephen J. Aves, Robert G. Beiko, Pedro Coutinho, Joel B. Dacks, Dion G. Durnford, Naomi M. Fast, Beverley R. Green, Cameron J. Grisdale, Franziska Hempel, Bernard Henrissat, Marc P. Höppner, Ken-Ichiro Ishida, Eunsoo Kim, Luděk Kořený, Peter G. Kroth, Yuan Liu, Shehre-Banoo Malik, Uwe G. Maier, Darcy McRose, Thomas Mock, Jonathan A. D. Neilson, Naoko T. Onodera, Anthony M. Poole, Ellen J. Pritham, Thomas A. Richards, Gabrielle Rocap, Scott W. Roy, Chihiro Sarai, Sarah Schaack, Shu Shirato, Claudio H. Slamovits, David F. Spencer, Shigekatsu Suzuki, Alexandra Z. Worden, Stefan Zauner, Kerrie Barry, Callum Bell, Arvind K. Bharti, John A. Crow, Jane Grimwood, Robin Kramer, Erika Lindquist, Susan Lucas, Asaf Salamov, Geoffrey I. McFadden, Christopher E. Lane, Patrick J. Keeling, Michael W. Gray, Igor V. Grigoriev, John M. Archibald
      Algal genomes reveal evolutionary mosaicism and the fate of nucleomorphs
      Nature, 492/2012, pages 59-65 DOI: 10.1038/nature11681chrome_reader_mode
    23. Richard G Dorrell, Gillian Gile, Giselle McCallum, Raphaël Méheust, Eric P Bapteste, Christen M Klinger, Loraine Brillet-Guéguen, Katalina D Freeman, Daniel J Richter, Chris Bowler
      Chimeric origins of ochrophytes and haptophytes revealed through an ancient plastid proteome
    24. Richard Wetherbee, Christopher J. Jackson, Sonja I. Repetti, Lesley A. Clementson, Joana F. Costa, Allison van de Meene, Simon Crawford, Heroen Verbruggen
      The golden paradox – a new heterokont lineage with chloroplasts surrounded by two membranes
      Journal of Phycology, 55/2019, pages 257-278 DOI: 10.1111/jpy.12822chrome_reader_mode
    25. Zoltán Füssy, Tereza Faitová, Miroslav Oborník
      Subcellular Compartments Interplay for Carbon and Nitrogen Allocation in Chromera velia and Vitrella brassicaformis
      Genome Biology and Evolution, 11/2019, pages 1765-1779 DOI: 10.1093/gbe/evz123chrome_reader_mode
    26. Olof Emanuelsson, Søren Brunak, Gunnar von Heijne, Henrik Nielsen
      Locating proteins in the cell using TargetP, SignalP and related tools
      Nature Protocols, 2/2007, pages 953-971 DOI: 10.1038/nprot.2007.131chrome_reader_mode
    27. Henrik Nielsen
      Predicting Secretory Proteins with SignalP
      Methods in Molecular Biology: Protein Function Prediction, 1611/2017, pages 59-73 DOI: 10.1007/978-1-4939-7015-5_6chrome_reader_mode
    28. Henrik Nielsen, Konstantinos D. Tsirigos, Søren Brunak, Gunnar von Heijne
      A Brief History of Protein Sorting Prediction
      The Protein Journal, 38/2019, pages 200-216 DOI: 10.1007/s10930-019-09838-3chrome_reader_mode
    29. José Juan Almagro Armenteros, Konstantinos D. Tsirigos, Casper Kaae Sønderby, Thomas Nordahl Petersen, Ole Winther, Søren Brunak, Gunnar von Heijne, Henrik Nielsen
      SignalP 5.0 improves signal peptide predictions using deep neural networks
      Nature Biotechnology, 37/2019, pages 420-423 DOI: 10.1038/s41587-019-0036-zchrome_reader_mode
    30. Jose Juan Almagro Armenteros, Marco Salvatore, Olof Emanuelsson, Ole Winther, Gunnar von Heijne, Arne Elofsson, Henrik Nielsen
      Detecting sequence signals in targeting peptides using deep learning
      Life Science Alliance, 2/2019, page e201900429 DOI: 10.26508/lsa.201900429chrome_reader_mode
    31. H Nielsen, J Engelbrecht, S Brunak, G von Heijne
      Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites.
      Protein Engineering, Design and Selection, 10/1997, pages 1-6 DOI: 10.1093/protein/10.1.1chrome_reader_mode
    32. Olof Emanuelsson, Henrik Nielsen, Søren Brunak, Gunnar von Heijne
      Predicting Subcellular Localization of Proteins Based on their N-terminal Amino Acid Sequence
      Journal of Molecular Biology, 300/2000, pages 1005-1016 DOI: 10.1006/jmbi.2000.3903chrome_reader_mode
    33. Weichao Huang, Carolina Río Bártulos, Peter G. Kroth
      Diatom Vacuolar 1,6‐β‐Transglycosylases can Functionally Complement the Respective Yeast Mutants
      Journal of Eukaryotic Microbiology, 63/2016, pages 536-546 DOI: 10.1111/jeu.12298chrome_reader_mode
    34. Viktoria Schreiber, Josefine Dersch, Katharina Puzik,more_horiz, Uwe G. Maier
      The Central Vacuole of the Diatom Phaeodactylum tricornutum: Identification of New Vacuolar Membrane Proteins and of a Functional Di-leucine-based Targeting Motif
    35. Weichao Huang, Ilka Haferkamp, Bernard Lepetit, Mariia Molchanova, Shengwei Hou, Wolfgang Jeblick, Carolina Río Bártulos, Peter G. Kroth
      Reduced vacuolar β-1,3-glucan synthesis affects carbohydrate metabolism as well as plastid homeostasis and structure in Phaeodactylum tricornutum
      Proceedings of the National Academy of Sciences, 115/2018, pages 4791-4796 DOI: 10.1073/pnas.1719274115chrome_reader_mode
    36. Carolina Río Bártulos, Matthew B Rogers, Tom A Williams, Eleni Gentekaki, Henner Brinkmann, Rüdiger Cerff, Marie-Françoise Liaud, Adrian B Hehl, Nigel R Yarlett, Ansgar Gruber, Peter G Kroth, Mark van der Giezen
      Mitochondrial Glycolysis in a Major Lineage of Eukaryotes
      Genome Biology and Evolution, 10/2018, pages 2310-2325 DOI: 10.1093/gbe/evy164chrome_reader_mode
    37. Alexander F Schober, Carolina Río Bártulos, Annsophie Bischoff, Bernard Lepetit, Ansgar Gruber, Peter G Kroth
      Organelle Studies and Proteome Analyses of Mitochondria and Plastids Fractions from the Diatom Thalassiosira pseudonana
      Plant and Cell Physiology, 60/2019, pages 1811-1828 DOI: 10.1093/pcp/pcz097chrome_reader_mode
    38. Pia Marter, Sebastian Schmidt, Stephan Kiontke, Daniel Moog
      Optimized mRuby3 is a Suitable Fluorescent Protein for in vivo Co-localization Studies with GFP in the Diatom Phaeodactylum tricornutum
    39. Jannick Dyrløv Bendtsen, Henrik Nielsen, Gunnarvon Heijne, Søren Brunak
      Improved Prediction of Signal Peptides: SignalP 3.0
      Journal of Molecular Biology, 340/2004, pages 783-795 DOI: 10.1016/j.jmb.2004.05.028chrome_reader_mode
    40. Henrik Nielsen, Anders Krogh
      Prediction of signal peptides and signal anchors by a hidden Markov model
      Proceedings International Conference on Intelligent Systems for Molecular Biology, 6/1998, pages 122-130 chrome_reader_mode
    41. Thomas Nordahl Petersen, Søren Brunak, Gunnar von Heijne, Henrik Nielsen
      SignalP 4.0: discriminating signal peptides from transmembrane regions
      Nature Methods, 8/2011, pages 785-786 DOI: 10.1038/nmeth.1701chrome_reader_mode
    42. Pierre Baldi, Søren Brunak, Yves Chauvin, Claus A. F. Andersen, Henrik Nielsen
      Assessing the accuracy of prediction algorithms for classification: an overview
      Bioinformatics, 16/2000, pages 412-424 DOI: 10.1093/bioinformatics/16.5.412chrome_reader_mode
    Commentslink

    Create a Matters account to leave a comment.