Proteogenomics Integrating Novel Junction Peptide Identification Strategy Discovers Three Novel Protein Isoforms of Human NHSL1 and EEF1B2

Abstract

In eukaryotes, alternative pre-mRNA splicing allows a single gene to encode different protein isoforms that function in many biological processes, and they are used as biomarkers or therapeutic targets for diseases. Although protein isoforms in the human genome are well annotated, we speculate that some low-abundance protein isoforms may still be under-annotated because most genes have a primary coding product and alternative protein isoforms tend to be under-expressed. A peptide coencoded by a novel exon and an annotated exon separated by an intron is known as a novel junction peptide. In the absence of known transcripts and homologous proteins, traditional whole-genome six-frame translation-based proteogenomics cannot identify novel junction peptides, and it cannot capture novel alternative splice sites. In this article, we first propose a strategy and tool for identifying novel junction peptides, called CJunction, which we then integrate into a proteogenomics process specifically designed for novel protein isoform discovery and apply to the analysis of a deep-coverage HeLa mass spectrometry data set with identifier PXD004452 in ProteomeXchange. We succeeded in identifying and validating three novel protein isoforms of two functionally important genes, NHSL1 (causative gene of Nance-Horan syndrome) and EEF1B2 (translation elongation factor), which validate our hypothesis. These novel protein isoforms have significant sequence differences from the annotated gene-coding products introduced by the novel N-terminal, suggesting that they may play importantly different functions.

Novel Proteoform Discovery by Precise Semi-De Novo Sequencing of Novel Junction Peptides

“…This resulted in a total of 958,121 theoretical novel junction peptides with lengths ≥ 7 (excluding “X”) and masses < 3000, of which 535,766 had “X” at the N-terminus, while 422,355 had “X” at the C-terminus. The method of constructing a theoretical database of novel junction peptides has also been described in detail in our previous work …”

“…The method of constructing a theoretical database of novel junction peptides has also been described in detail in our previous work. 22 SpliceNovo first searches the MS/MS spectra against the theoretical novel junction peptide database and screens the top N (e.g., 20) novel junction peptides with unknown "X" at the N-terminus or C-terminus (Figure 1a). When calculating the match scores of incomplete novel junction peptides to the spectra, 21 types of fragment ions were considered (b, y, a, b(2+), y(2+), a(2+), b-H2O, y-H2O, a-H2O, b-NH3, y-NH3, a-NH3, and by-type internal ions between 2 and 10 in length).…”

“…Especially for spectra with poor fragment ion coverage, the best methods are now not as accurate as database searching . We have also previously proposed a method, CJunction, to enumerate all possible novel junction peptides in MS data, but this method is based on the results of PEAKS and does not change the accuracy of the spectrum sequencing …”

“…21 We have also previously proposed a method, CJunction, to enumerate all possible novel junction peptides in MS data, but this method is based on the results of PEAKS and does not change the accuracy of the spectrum sequencing. 22 Based on the specificity in the composition of the novel junction peptides, we developed the semi-de novo sequencing algorithm SpliceNovo for identifying the novel junction peptides. SpliceNovo can significantly improve the accuracy of junction peptide identification.…”

Deep learning methods for de novo peptide sequencing

“…Evaluating the tools on human and mouse antibody data, the authors concluded that Casanovo and PointNovo show improved peptide recall across different enzymes and datasets compared with competing methods. Second, Tran et al [76] have evaluated PEAKS, PointNovo, Casanovo, and GraphNovo on five datasets: human tryptic data, Method Year Application Ind DeepNovo 2020 Detection of neoantigens [77] 2021 De novo sequencing in metaproteomics [78] ✓ 2021 Detection of neoantigens [79] ✓ 2021 Detection of junction peptides [80] ✓ 2021 Detection of shell proteins [81] ✓ 2023 Detection of neuropeptides [82] ✓ 2023 Detection of proteasome-generated spliced and non-spliced peptides [83] ✓ 2023 Antibody sequencing [84] ✓ 2023 Detection of short peptides [85] ✓ 2023 Noncanonical antigen detection [86] ✓ 2023 Detection of neoantigens [87] 2024 Antibody sequencing [88] ✓ 2024 Detection of neoantigens [89] 2024 Detection of venom proteins [ ✓ 2022 Detection of neoantigens [103] ✓ Table 3: Applications of deep learning de novo sequencing methods. The final column ("Ind") indicates whether the application was published independently of the original authors of the work.…”