Super-RNova
A transformer-based framework identifies peptides and unexpected posttranslational modifications directly from tandem mass spectra
A transformer-based de novo peptide sequencing framework has been developed to identify peptides and unknown or unexpected posttranslational modifications directly from tandem mass spectrometry data, without relying on a protein database or predefined modification list.
The method, called RNovA, treats each MS/MS spectrum as a graph of fragment ion peaks and divides the sequencing task between two modules: PathSearcher and SeqFiller. PathSearcher identifies likely fragmentation paths and unresolved mass gaps that may represent missing residues or unknown modifications, while SeqFiller reconstructs full peptide sequences from the same spectral evidence. By encoding mass differences between peaks, the framework can search a more flexible mass space than conventional residue-by-residue prediction.
The team designed RNovA for “open” posttranslational modification discovery. Recurring mass tags extracted from spectra are clustered into candidate modifications and passed to SeqFiller for sequence reconstruction, allowing modified peptides to be searched without retraining the model or supplying a fixed list of candidate residues.
In benchmark tests, RNovA achieved state-of-the-art de novo sequencing performance on standard datasets and outperformed PEAKS on synthetic peptides carrying diverse posttranslational modifications, including several modification types absent from its training data. In a spiked-peptide experiment, it recovered four artificial modifications from LC–MS/MS data acquired in a human cell digest background, while an open-search workflow detected only one.
The team then applied RNovA to real biological samples, including rheumatoid arthritis proteomics data, where it identified kynurenine-modified peptides that were validated using synthetic reference peptides. In a bacterial strain lacking a reference proteome, it also detected an unannotated glutamic acid modification.
A Molecular Split in the Po Valley
Molecular profiling of PM2.5 revealed site-specific seasonal chemistry that mass-based air-quality metrics would have missed
A year-long molecular survey of fine particulate matter in Italy’s Po Valley has shown that urban and agricultural aerosols can differ sharply in organic composition, even when bulk pollution measurements appear similar.
The team analyzed PM2.5 filter samples collected in 2021 from Milan and Schivenoglia, a rural agricultural site about 150 km away. Organic aerosol extracts were separated by ultra-high-performance liquid chromatography and analyzed by heated electrospray ionization high-resolution Orbitrap mass spectrometry in positive and negative ion modes. The resulting non-target dataset was processed through molecular formula assignment, compound-family grouping, and hierarchical clustering of time-series patterns.
That workflow detected thousands of organic features and separated them into seasonal clusters with source-related patterns. At both sites, summer clusters were dominated by CHO- and CHOS-containing compounds consistent with oxidation products of biogenic emissions, while winter clusters were enriched in nitrogen-containing compounds linked to combustion sources. Some wintertime molecules had high conjugation and were classified as potential contributors to brown carbon, the light-absorbing fraction of organic aerosol.
The non-target analysis also revealed site-specific chemistry that bulk measurements would have missed. At the agricultural site, plant-protection products made a clear contribution to the molecular profile, with the highest pesticide-related intensities appearing in May, when PM2.5 and organic carbon concentrations were at their lowest. At the urban site, site-specific summer clusters contained more nonpolar high-molecular-mass compounds and CHO-containing aromatics.
The authors suggest that the results expose a gap in mass-based air-quality assessment, showing that chemically distinct organic aerosol mixtures, including source-specific trace compounds, can emerge even when PM2.5 levels appear relatively low.
Paleoclimate at Micrometer Resolution
FT-ICR MSI assesses how much climate signal survives at micrometer-scale resolution in laminated and mixed marine sediments
Mass spectrometry imaging of marine sediments has helped test how much climate signal can be recovered from alkenone-based sea surface temperature records at micrometer-scale resolution, where biomarker heterogeneity and sediment mixing can limit record reliability.
The study analyzed Santa Barbara Basin sediment cores, where laminated and mixed intervals preserve different degrees of depositional structure. Using 100 µm cryomicrotome slices, the team mapped alkenone biomarkers with laser desorption ionization Fourier-transform ion cyclotron resonance mass spectrometry imaging. These algal lipid biomarkers are widely used to reconstruct past sea surface temperatures; the team also tracked pyropheophorbide α, a chlorophyll breakdown product used as a broad productivity marker.
Rather than assuming finer spatial resolution would automatically improve the record, the researchers used replicated measurements to test how reproducible the proxy signal was across nearby sediment slices. The alkenone maps showed strong spatial heterogeneity, including small clusters of similar proxy values within laminae. Replicate maps had highly consistent mean values, but time series built from individual 200 µm sediment horizons agreed less well at the finest scales, reflecting noise, local heterogeneity, or slight depth-alignment differences.
Variogram and signal-to-noise analyses showed that sediment structure controlled how much proxy variability could be interpreted. Well-preserved laminated intervals showed stronger within-horizon similarity and higher replicability, while mixed intervals had reduced signal content. Across the Santa Barbara Basin material, signal-to-noise ratios increased with temporal aggregation, from roughly interannual toward subdecadal timescales.
The authors suggest that the approach could guide future MSI-based paleoclimate studies by identifying when finer sampling adds meaningful climate information and when replication, averaging, or uncertainty estimates are needed to keep sediment noise in check.
A Clearer Glycan Clock for Aging
Absolute MALDI-TOF measurements of IgG glycans were used to build a biological-age model that tracks aging in mice
A study of aging in mice has used absolute measurements of immunoglobulin G (IgG) glycans to build a biological-age model, suggesting that changes in antibody glycosylation can be tracked more directly than by relative glycan abundance alone.
The team followed laboratory mice across seven time points from adulthood to old age, profiling IgG N-glycans by MALDI-TOF mass spectrometry. Instead of measuring each glycan only as a proportion of the total glycan pool, the workflow used a deuterated internal full-glycome standard designed to mirror the mouse IgG glycan profile, together with external standards for selected aging-linked glycans. This allowed the researchers to convert mass spectrometry peak signals into absolute glycan concentrations.
Two glycan patterns stood out. A bisected IgG glycan decreased with age, while a digalactosylated glycan increased, changes that were partly reversed in mice kept under caloric restriction. Using the absolute concentrations of these two markers, the team built the abGlycoAge model, which tracked chronological age in naturally aging mice and predicted a younger biological state in caloric restriction cohorts.
To connect those glycan shifts with possible biology, the researchers paired the glycomic data with RNA sequencing of splenic B cells. The transcriptomic results pointed to altered N-linked glycosylation pathways during aging, with some age-associated gene-expression changes partly reversed by caloric restriction. The study also included a small preliminary test of IgG modified with younger mice, which was associated with reduced inflammatory markers and lower senescence-associated β-galactosidase staining in selected tissues at the higher dose.
(Mass) Spectacular and Strange
“His Breath is Worse Than His Bite”
Dogs have a strong claim to the title of “man’s best friend”: loyal, affectionate, and emotionally intelligent. Unfortunately, they can also smell as if something small has died behind their molars. A new study, however, suggests that sugar cane molasses, or more specifically the polyphenols extracted from it, could help.
Researchers tested a sugar cane polyphenol oral spray in 10 pet dogs with noticeable halitosis, combining human sensory scoring with saliva volatile profiling and microbiome analysis. In the short-term trial, trained panelists judged oral odor before and after spraying, while saliva samples were analyzed by solid-phase microextraction gas chromatography–mass spectrometry (SPME-GC-MS) to track odor-associated compounds.
The spray reduced perceived malodor within the first hour, while SPME-GC-MS showed decreases in several volatile compounds linked to unpleasant breath, including indole, amines, aldehydes, and ester-like odorants. After 30 days, volatile profiling showed further reductions in odor-related compounds, alongside reduced levels of bacteria associated with foul-smelling metabolism.
“Importantly, the deodorizing effect was not simply due to odor masking, because the GC-MS results showed reductions in several odor-associated compounds in saliva,” said lead author Hongye Li in the team’s press release.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
