QuantFusion: Novel Unified Methodology for Enhanced Coverage and Precision in Quantifying Global Proteomic Changes in Whole Tissues

Single quantitative platforms such as label-based or label-free quantitation (LFQ) present compromises in accuracy, precision, protein sequence coverage, and speed of quantifiable proteomic measurements. To maximize the quantitative precision and the number of quantifiable proteins or the quantifiable coverage of tissue proteomes, we have developed a unified approach, termed QuantFusion, that combines the quantitative ratios of all peptides measured by both LFQ and label-based methodologies. Here, we demonstrate the use of QuantFusion in determining the proteins differentially expressed in a pair of patient-derived tumor xenografts (PDXs) representing two major breast cancer (BC) subtypes, basal and luminal. Label-based in-spectra quantitative peptides derived from amino acid-coded tagging (AACT, also known as SILAC) of a non-malignant mammary cell line were uniformly added to each xenograft with a constant predefined ratio, from which Ratio-of-Ratio estimates were obtained for the label-free peptides paired with AACT peptides in each PDX tumor. A mixed model statistical analysis was used to determine global differential protein expression by combining complementary quantifiable peptide ratios measured by LFQ and Ratio-of-Ratios, respectively. With minimum number of replicates required for obtaining the statistically significant ratios, QuantFusion uses the distinct mechanisms to “rescue” the missing data inherent to both LFQ and label-based quantitation. Combined quantifiable peptide data from both quantitative schemes increased the overall number of peptide level measurements and protein level estimates. In our analysis of the PDX tumor proteomes, QuantFusion increased the number of distinct peptide ratios by 65%, representing differentially expressed proteins between the BC subtypes. This quantifiable coverage improvement, in turn, not only increased the number of measurable protein fold-changes by 8% but also increased the average precision of quantitative estimates by 181% so that some BC subtypically expressed proteins were rescued by QuantFusion. Thus, incorporating data from multiple quantitative approaches while accounting for measurement variability at both the peptide and global protein levels make QuantFusion unique for obtaining increased coverage and quantitative precision for tissue proteomes. Single quantitative platforms such as label-based or label-free quantitation (LFQ) present compromises in accuracy, precision, protein sequence coverage, and speed of quantifiable proteomic measurements. To maximize the quantitative precision and the number of quantifiable proteins or the quantifiable coverage of tissue proteomes, we have developed a unified approach, termed QuantFusion, that combines the quantitative ratios of all peptides measured by both LFQ and label-based methodologies. Here, we demonstrate the use of QuantFusion in determining the proteins differentially expressed in a pair of patient-derived tumor xenografts (PDXs) representing two major breast cancer (BC) subtypes, basal and luminal. Label-based in-spectra quantitative peptides derived from amino acid-coded tagging (AACT, also known as SILAC) of a non-malignant mammary cell line were uniformly added to each xenograft with a constant predefined ratio, from which Ratio-of-Ratio estimates were obtained for the label-free peptides paired with AACT peptides in each PDX tumor. A mixed model statistical analysis was used to determine global differential protein expression by combining complementary quantifiable peptide ratios measured by LFQ and Ratio-of-Ratios, respectively. With minimum number of replicates required for obtaining the statistically significant ratios, QuantFusion uses the distinct mechanisms to “rescue” the missing data inherent to both LFQ and label-based quantitation. Combined quantifiable peptide data from both quantitative schemes increased the overall number of peptide level measurements and protein level estimates. In our analysis of the PDX tumor proteomes, QuantFusion increased the number of distinct peptide ratios by 65%, representing differentially expressed proteins between the BC subtypes. This quantifiable coverage improvement, in turn, not only increased the number of measurable protein fold-changes by 8% but also increased the average precision of quantitative estimates by 181% so that some BC subtypically expressed proteins were rescued by QuantFusion. Thus, incorporating data from multiple quantitative approaches while accounting for measurement variability at both the peptide and global protein levels make QuantFusion unique for obtaining increased coverage and quantitative precision for tissue proteomes. The past decade has witnessed rapid progress in mass spectrometry (MS)-based quantitative proteomics with the development of software and data analysis tools to interrogate large amounts of MS data. Quantitative proteomic technologies have shown great potential in delineating dysregulated proteomes in diseases such as cancer (1.Chen E.I. Yates 3rd, J.R. Cancer proteomics by quantitative shotgun proteomics.Mol. Oncol. 2007; 1: 144-159Crossref PubMed Scopus (65) Google Scholar, 2.Cox J. Mann M. Quantitative, high-resolution proteomics for data-driven systems biology.Annu. Rev. Biochem. 2011; 80: 273-299Crossref PubMed Scopus (531) Google Scholar, 3.Chaerkady R. Pandey A. Quantitative proteomics for identification of cancer biomarkers.Proteomics Clin. Appl. 2007; 1: 1080-1089Crossref PubMed Scopus (29) Google Scholar, 4.Zhang B. Wang J. Wang X. Zhu J. Liu Q. Shi Z. Chambers M.C. Zimmerman L.J. Shaddox K.F. Kim S. Davies S.R. Wang S. Wang P. Kinsinger C.R. Rivers R.C. Rodriguez H. Townsend R.R. Ellis M.J. Carr S.A. Tabb D.L. Coffey R.J. Slebos R.J. Liebler D.C. and NCI CPTAC. Proteogenomic characterization of human colon and rectal cancer.Nature. 2014; 513: 382-387Crossref PubMed Scopus (937) Google Scholar). Quantitative schemes via either stable isotope labeling or label-free quantitation (LFQ) 1The abbreviations used are:LFQlabel-free quantitationRoRRatio-of-RatioBCbreast cancerPDXpatient-derived tumor xenograftFDRfalse discovery rateAACTamino acid-coded taggingbRPLCbasic reversed phase chromatographyLHlight to heavy. are used widely to assist MS for quantitative assessments of the changes in protein expression, post-translational modifications (5.Zhu H. Hunter T.C. Pan S. Yau P.M. Bradbury E.M. Chen X. Residue-specific mass signatures for the efficient detection of protein modifications by mass spectrometry.Anal. Chem. 2002; 74: 1687-1694Crossref PubMed Scopus (55) Google Scholar), and protein-protein interactions (6.Wang T. Gu S. Ronni T. Du Y.C. Chen X. In vivo dual-tagging proteomic approach in studying signaling pathways in immune response.J. Proteome Res. 2005; 4: 941-949Crossref PubMed Scopus (50) Google Scholar) in many biological systems, including tumor samples (7.Neilson K.A. Ali N.A. Muralidharan S. Mirzaei M. Mariani M. Assadourian G. Lee A. van Sluyter S.C. Haynes P.A. Less label, more free: approaches in label-free quantitative mass spectrometry.Proteomics. 2011; 11: 535-553Crossref PubMed Scopus (537) Google Scholar, 8.Eberl H.C. Spruijt C.G. Kelstrup C.D. Vermeulen M. Mann M. A map of general and specialized chromatin readers in mouse tissues generated by label-free interaction proteomics.Mol. Cell. 2013; 49: 368-378Abstract Full Text Full Text PDF PubMed Scopus (134) Google Scholar, 9.Choi H. Liu G. Mellacheruvu D. Tyers M. Gingras A.C. Nesvizhskii A.I. Analyzing protein-protein interactions from affinity purification-mass spectrometry data with SAINT.Curr. Protoc. Bioinformatics. 2012; (Chapter 8, Unit 8.15)Crossref Scopus (95) Google Scholar, 10.Bantscheff M. Lemeer S. Savitski M.M. Kuster B. Quantitative mass spectrometry in proteomics: critical review update from 2007 to the present.Anal. Bioanal. Chem. 2012; 404: 939-965Crossref PubMed Scopus (581) Google Scholar, 11.Navarro P. Trevisan-Herraz M. Bonzon-Kulichenko E. Núñez E. Martínez-Acedo P. D. R. E. M. J. J. statistical for quantitative proteomics by stable isotope Proteome Res. 2014; PubMed Scopus Google Scholar). the of accuracy, and in the analysis of from with the quantitative to for quantitative mass spectrometry proteomic with Bioinformatics. 2012; PubMed Google Scholar) for tumor analysis have the of mass such as mass or isotope tagging for and quantitation the of with Chem. 2012; PubMed Scopus Google Scholar, P. Liu T. R.J. Townsend R. P. Davies S.R. D. S. Rodriguez H. Liebler D. Ellis M. Carr S.A. in and changes in pathways but not global protein Cell. 2014; Full Text Full Text PDF PubMed Scopus Google Scholar). for quantitative analysis of large mass and isotope tagging for and quantitation are to the of large amounts of protein as The use of added peptide derived from cell or labeling to peptides Liu T. A. R.J. and the to quantitative discovery proteomics by the use of Proteome Res. PubMed Scopus (50) Google Scholar) for quantitation of expression in both and MS for quantitation of The of in-spectra quantitative cell X. Bradbury E.M. mass tagging with stable in proteins for and efficient protein Chem. PubMed Scopus Google Scholar, H. Pan S. Gu S. Bradbury E.M. Chen X. stable isotope labeling for quantitative 2002; PubMed Scopus Google Scholar), in vivo quantitation amino acid-coded (AACT, also known as or stable isotope labeling by amino in cell B. H. Pandey A. Mann M. isotope labeling by amino in cell as a and approach to expression proteomics.Mol. Cell. 2002; 1: Full Text Full Text PDF PubMed Scopus Google the for quantitation of changes in protein biological for tissue a cell line as M. Mann M. in tissues by a 2011; PubMed Scopus Google Scholar) a of cell T. J. P. J.R. Mann M. for quantitative proteomics of human tumor PubMed Scopus Google to a to peptides that are either missing or present at The missing that to to as quantitative of tissue that has by the of peptide S.A. of quantitative for in Chem. 2013; PubMed Scopus Google Scholar). A more labeling such as labeling of the tissue of the of via stable isotope labeling of has Yates 3rd, J.R. isotope labeling of for in vivo quantitative proteomic 2013; PubMed Scopus Google Scholar, T. B. Yates J.R. proteomic analysis of tissues 2011; PubMed Scopus Google Scholar). The and with and labeling use of label-free quantitation Ratio-of-Ratio breast cancer patient-derived tumor xenograft discovery amino acid-coded tagging reversed phase to heavy. quantitation of tissue and tumor proteins to LFQ and has as to A. A of labeling and label-free mass proteomics Proteome Res. PubMed Scopus Google Scholar, A. J.R. K.A. of label-free for human proteins by shotgun proteomics.Mol. Cell. 2005; 4: Full Text Full Text PDF PubMed Scopus Google Scholar). the inherent precision and of LFQ multiple or as to LFQ some each a number of peptide LFQ inherent to such as or labeling in and the for the samples to of to D. J. for in quantitative 2007; 4: PubMed Scopus Google Scholar). LFQ the as as with label-based quantitative such as AACT that the B. H. Pandey A. Mann M. isotope labeling by amino in cell as a and approach to expression proteomics.Mol. Cell. 2002; 1: Full Text Full Text PDF PubMed Scopus Google Scholar). that the of multiple quantitative schemes and to more determine the changes of protein with a coverage of tissue tumor subtypes. combining peptide both LFQ and AACT Ratio-of-Ratio the overall number of quantifiable changes to a of peptide and between two LFQ to the of the complementary peptide LFQ quantitation ratios between samples of a also at of the samples a peptide the label-based To a complementary quantitative we our development of a unified quantitative approach, termed QuantFusion, that uses a mixed model to interrogate quantifiable peptide data derived from both LFQ and label-based AACT from a MS LFQ and measurements complementary and to the number of replicates required for the statistically significant LFQ The of combining with and each protein the use of a statistical demonstrate the of the mixed approach the of the in two major breast cancer (BC) subtypes. QuantFusion increased by the number of distinct peptide ratios to This of quantifiable peptide coverage, in turn, increased the number of measurable protein fold-changes by 8% and increased the average precision of quantitative peptide estimates by The used to the statistical model with a data used in are to to QuantFusion xenograft breast were Ellis M.J. S. Chen J. D.C. H. J. Q. Chen M.C. E. Davies S. T. R. K.A. D. Du M. G. R. M. in a breast cancer and PubMed Scopus Google Scholar, S. D. J. R. Liu A. X. Liu S. J. D. M. T. J. R. T. A. Davies S.R. J. D. T. M. R. R. R. R. Wang S. A. J.R. Ellis M.J. by characterization of 2013; 4: Full Text Full Text PDF PubMed Scopus Google Scholar) and to B. Wang J. Wang X. Zhu J. Liu Q. Shi Z. Chambers M.C. Zimmerman L.J. Shaddox K.F. Kim S. Davies S.R. Wang S. Wang P. Kinsinger C.R. Rivers R.C. Rodriguez H. Townsend R.R. Ellis M.J. Carr S.A. Tabb D.L. Coffey R.J. Slebos R.J. Liebler D.C. and NCI CPTAC. Proteogenomic characterization of human colon and rectal cancer.Nature. 2014; 513: 382-387Crossref PubMed Scopus (937) Google Scholar, P. Liu T. R.J. Townsend R. P. Davies S.R. D. S. Rodriguez H. Liebler D. Ellis M. Carr S.A. in and changes in pathways but not global protein Cell. 2014; Full Text Full Text PDF PubMed Scopus Google Scholar). The were to and protein a of and of protein was with and with The proteins were to with for at The was with and with for at was by the of to and the was by The peptide were and in a for reversed phase A was by from to A was and was in A of were for each of the basal and samples and to The were and a were to the and in analysis was via reversed phase a to a mass The was to peptides a at that was from the via a was by the to the in line with a of all the peptides was with a of at a A was and was in were also in a with MS to a mass of and a of at by of the was used to peptides at a of in the of were derived from replicates of each tumor and were to two in the of for global peptide were and peptide identification was the in software protein were the human and mouse protein sequence from the M.J. M. Carr S.A. Townsend R.R. Kinsinger M. Rodriguez H. Liebler D.C. to cancer with proteomics: the NCI 2013; PubMed Scopus Google Scholar). This and the were derived from the were with a approach a of protein of the and of The of proteins was obtained from the Proteome was with a discovery of and peptides were to proteins with a protein of A mass of was used for the that for of that were to a a mass of and a mass to two at the of and protein as a of was as a are by of and and to of the mass spectrometry data PDX tumor samples were at the as and for LFQ was The measured the of and the of a peptide were via the label-free quantitation in J. Mann M. label-free by and peptide termed Cell. 2014; Full Text Full Text PDF PubMed Scopus Google Scholar). replicates for each PDX were in the LFQ with quantitation unique and peptide to with a of peptide of and protein of The peptide and protein were and in and statistical model and of a of data were and that was developed of the are in and peptide are in with in the proteomic used for obtaining quantitative of the two PDX tumor and with of non-malignant mammary cell line used for in-spectra The cell line was in a with and or The with protein from the PDX by of proteins from the tissue tumor protein at a The peptide were to and to To the of the peptides in the protein was by a the of were and that for the of the and detection of peptides by the the are at the global peptide and each was either LFQ of both samples or by obtaining AACT from the added in and the peptide tissue that both measurements by LFQ and were obtained from the of each for with added in and with with added in respectively. at the peptide level with the that identification and are was the added in and the added in with two replicates respectively. was used to both the data with LFQ analysis and quantitative analysis in for each were and the were that many to LFQ data and analysis of the of used the ratios of the peptide level measurements as by but our model of LFQ ratios to a analysis only peptide ratios are LFQ a peptide used as a of the protein a protein from to in and from to in required for a protein Thus, our analysis was only LFQ that have in both samples and to ratios that to proteins that In a in that to a protein with in With the data we the ratios and each a are ratios that In protein each LFQ peptide the protein ratio, and we and protein ratios between the and the protein in and respectively. the model we that This to the model for the ratios, the LFQ or of the peptide the so that the in for the peptide from protein a in and LFQ the LFQ peptide from to a representing the average all proteins T. a that the to in for protein a representing the for protein between and the of variability and and of which the variability in proteomics to at the average fold-changes all This has by as a In the and that have the protein from for the and LFQ respectively. that the of by the of so that peptide peptide with the obtained from our LFQ a to the have to and the LFQ ratios have to only of a In the protein from both the LFQ ratios and the we to as our we have the of our estimates of are in the of Scholar). This approach A model have to the analysis as was by and for quantitative mass spectrometry proteomic with Bioinformatics. 2012; PubMed Google Scholar). to all the data at The also some a protein from a number of peptide ratios, the the average of all protein This and in for PubMed Scopus Google Scholar). that from the inherent to mass spectrometry and the potential for are by the estimates. were for each protein and the of the was as To proteins of we the estimates by estimates. are to in a and the used to as by The discovery A and the Scopus Google Scholar). proteins with a of were for the model and data are in the To peptide for quantitation either LFQ or used we developed a statistical termed QuantFusion, to data obtained via each a MS data from two as label-free peptide of both tumor and ratios derived the peptide of the cell that the use of the peptides at a quantitative in both the of peptides from the in MS data are from either the or of the peptide ratios J. Mann M. peptide identification mass and protein PubMed Scopus Google Scholar) of from each amino sequence has a of average of the number of that make a mass but to make protein that the are the for peptides a the in a a of the in each that the are the for all peptide ratios to the protein in the the protein by the average peptide statistical that estimates to the average for quantitative mass spectrometry proteomic with Bioinformatics. 2012; PubMed Google Scholar, J. H. A. R. P. J. expression for label-free quantitative Bioinformatics. 2012; PubMed Scopus Google Scholar, J. T. J. H. A statistical for protein quantitation in PubMed Scopus Google Scholar). not to missing and as a model estimates of protein changes to the The to the model approach in the of a obtained from to for missing data a but not that we our a to the number of missing In the LFQ a peptide missing a measured peptide in A has in B. in the a peptide missing the in A to the peptide also in A. missing data mechanisms are which that we both quantitative schemes we the of peptide ratios in either for use in our combining both some the we all of the data for a have measurements for the peptide in samples A and and respectively. also have the and and to the LFQ J. Mann M. label-free by and peptide termed Cell. 2014; Full Text Full Text PDF PubMed Scopus Google Scholar), which LFQ and Thus, ratios as the LFQ the and the The for a peptide as which the peptides were mixed in in each at to average the and LFQ such a significant for two the are not and from a of and which that as two of data to used in the protein to estimates of the have the of the LFQ estimates in the make more to To for the of we have developed a statistical model that estimates and the of the data. This model in and the of all global peptides for and that the LFQ and AACT are by and the protein of between and tumor The shown for each with representing the we obtained the AACT at for global and by combining both ratios, we obtained estimates that in the expression of proteins between and with from the cell line as the the overall number of peptides and proteins and the QuantFusion The number of quantifiable peptides was and for and QuantFusion. The number of quantifiable protein was and for and QuantFusion proteins are in This the of the unified approach to the number of of quantifiable peptides to the also a of the number of LFQ and protein in the with analysis that used only LFQ the QuantFusion increased by the number of unique peptides used for measurements of protein in turn, to 8% in the number of unique proteins the average precision was increased by the between LFQ and for the global protein with a of The protein are not in the of a missing either obtained by or obtained by such as A. E. A. A. J. Mann M. S. proteomics a mass Cell. 2011; Full Text Full Text PDF PubMed Scopus Google Scholar) or A. M. M.J. for 2013; PubMed Scopus Google Scholar). QuantFusion to as a unified present in either LFQ or The unified for all missing in either LFQ or are in significant proteins a of This at of or to the of of both and the variability of the precision with the two In we in the the was representing the that a are in and both are are in in the and for and QuantFusion. The and significant protein that are differentially expressed between and by and QuantFusion, for a of the to the two which use only LFQ or the QuantFusion the number of significant fold-changes by from that obtained by The were obtained a to significant changes between and The and significant protein differentially expressed between and in and QuantFusion, the that a The in the number of significant protein between and the QuantFusion for discovery proteomics as the protein for biological or In to the quantitative coverage, QuantFusion and used to the precision obtained with each In our we two replicates from each To as shown the of we each with LFQ and QuantFusion, a of of protein estimates. The was by the and the average of The was and for and QuantFusion, respectively. The the that a of the of protein estimates The between two was and for and QuantFusion, respectively. demonstrate that QuantFusion the and the precision demonstrate the precision of QuantFusion This not the QuantFusion from a of that the from the data of quantifiable peptides used in QuantFusion analysis the of incorporating the model the LFQ of QuantFusion of fold-changes the of the precision estimates the and have the precision and estimates. significant proteins both the of a and the precision of the This the from two for significant The for a The the that the in the the protein that are significant to both are in that are significant are in and that are significant are in were not significant either of are as of that both and and The of that each protein the Here, we multiple to the data. the we have the from both replicates with each of the we are the data and each This to the for each used to the precision obtained with each This was by the and the average of The the that a of the of protein estimates LFQ QuantFusion that our demonstrate that the QuantFusion both the and the that the has the which that the from the data used in the QuantFusion the of incorporating the model the LFQ QuantFusion to analysis LFQ the are estimates QuantFusion and LFQ The estimates are between with a of QuantFusion the by we that the by This was by estimates from the LFQ model with both replicates and to estimates from the LFQ model that only The obtained in are in for and QuantFusion are and respectively. for all are in our that the estimates the but the LFQ estimates in The of the number of estimates and variability has a the biological of protein The between and in the between and that in with to we The average proteins was for each model and used as the to as the of the of our protein and the of the the the as shown in for a as the of or were the of in was to that has of and fold-changes of and are to have a of the with and QuantFusion, respectively. the protein estimates obtained by with estimates obtained by QuantFusion. QuantFusion the number of the quantifiable proteins that the statistically significant BC changes in which in the of a coverage for the biological between basal and breast cancer shown in the in QuantFusion differentially expressed QuantFusion, proteins that were not differentially expressed in a not quantifiable or in LFQ were to differentially expressed proteins were to differentially expressed proteins that were not differentially expressed not quantifiable or in both LFQ and proteins were to differentially expressed were rescued proteins are in the QuantFusion of A at the that in of proteins with increased and differential expression LFQ the from QuantFusion. A of proteins in and LFQ and we a number of proteins in that are to with the more of breast protein with tumor M. D. M. S. G. P.A. B. van B. cell as breast cancer Clin. PubMed Scopus Google Scholar). both LFQ and the for are not With QuantFusion, the protein a of of breast cancer Chen of for in breast cancer in 2011; PubMed Scopus Google Scholar) and was not differentially expressed in both LFQ and but was in QuantFusion. and of was not differentially expressed in LFQ but was in QuantFusion and signaling has shown to a distinct in basal analysis a of expression to breast in human 2014; PubMed Scopus Google Scholar). protein from both LFQ and has shown to and to in cancer Liu Wang M. Chen Q. B. Wang Wang J. J. J. of by the in cancer 2014; PubMed Scopus Google Scholar). a in a for to a in basal breast Here, we present the development of a unified quantitative approach, QuantFusion, for the number of peptides to determine global protein expression between approach combines of via LFQ and ratios between tumor peptides and stable peptide QuantFusion the quantitative of between tumor by obtaining have all data with a model that missing data in either LFQ or for more unified estimates for protein expression The and the data approach we have used for combining two quantitative measurements a unified we QuantFusion more significant of global protein expression in and PDX tumor with each LFQ or the missing present in both LFQ and have not from or of the unified approach the missing to some by all the data from both in global protein expression between the QuantFusion increased by the number of distinct peptide ratios used in our This in turn, increased the number of measurable protein fold-changes by 8% and increased the average precision of our estimates by The used to the statistical model with data used in are to to QuantFusion In QuantFusion for of proteomic such as affinity and label-free interaction proteomics in the to The use of QuantFusion in proteomic in a

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.