Skip to main content
Fig. 4 | Molecular Neurodegeneration

Fig. 4

From: Deep proteomic network analysis of Alzheimer’s disease brain reveals alterations in RNA binding proteins and RNA splicing associated with disease

Fig. 4

Alternative Splicing in AD. ae Workflow for proteogenomic analysis of alternative splicing (a). RNA-Seq is performed on mRNA isolated from dorsolateral prefrontal cortex (DLPFC) control and AD brain. The mRNA sequences are then translated and digested with a given enzyme in silico to obtain peptide sequences. Peptide sequences that contain non-canonical exon-exon junctions (alt-EEjxns) are appended to the search database for peptide identification and subsequent quantification. b Illustration of alt-EEjxn peptide quantification for the clathrin light chain B protein (CLTB). An alternative splicing event between exons 5, 6, and 9 leads to the generation two alt-EEjxn peptides after enzymatic digestion. The levels of these alt-EEjxn peptides, each of which reflects a particular splicing “decision,” can be quantified across case groups. The exon numbering shown in panel (b) is based on the Gencode v19 exon annotation database, and includes RNA-derived junctions. c Venn diagram representing the number of alt-EEjxn peptides quantifiable by LFQ-trypsin or TMT-LysC pipelines in the BLSA cohort. A peptide was considered quantifiable if it had at least two measurements in at least two different case groups. The overlap between TMT-LysC and LFQ-trypsin represents quantifiable exon-exon junctions that were identified by both methods, even though the peptides that contain the junction may be different between the two methods. d Differential abundance of alt-EEjxns between AsymAD and AD, color-coded by the module to which each junction peptide is most highly correlated. The amino acid sequence of each alt-EEjxn peptide, and the module to which each alt-EEjxn peptide is most highly correlated, is provided in interactive HTML files for each case group comparison in Supplementary Data. e Enrichment of TMT-LysC alt-EEjxn peptides in TMT-LysC network modules. Each network module eigenprotein was correlated with molecular phenotypes obtained from the same tissue sample as the alt-EEjxn peptides, as well as with cognition as measured by the last MMSE score proximate to death. Modules within the dashed box are unique to the TMT network, similar to the depiction shown in Fig. 1a. Module numbers highlighted in red indicate modules enriched in RNA binding proteins as shown in Additional file 17: Figure S10. Significance of enrichment for alt-EEjxn peptides was calculated by Fisher exact test, and is shown by single color heat map of -log10 p value (increased red represents smaller p value). P values are corrected by Benjamini-Hochberg FDR. Module eigenprotein correlation with molecular species and cognition is shown by two-color heat map (red represents positive correlation, blue represents negative correlation). MMSE correlation was calculated by Spearman test. All other correlations are bicorrelation rho. Aβ, amyloid-β(17–28); MAPT pT231, microtubule associated protein tau peptide (VAVVRpTPPKSPSSAK) phosphorylated at threonine 231; U1-70K, U1 small nuclear ribonucleoprotein 70 kDa; TDP-43, TAR DNA-binding protein 43; MMSE, mini-mental state examination. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Back to article page