A Larger Empirical Library for Metabolomics 

METLIN 960 K adds standardized empirical MS/MS spectra for more than 960,000 small-molecule standards across untargeted and targeted workflows

A major expansion of METLIN has added standardized empirical MS/MS data for more than 960,000 small-molecule standards, supporting spectral matching in untargeted and targeted workflows. 

The study, from Gary Siuzdak’s group at Scripps Research, describes METLIN 960 K as an empirical resource based on authentic molecular standards rather than predicted or community-contributed spectra. Each compound has high-resolution MS/MS spectra in positive and negative ionization modes across four collision energies, typically 0, 10, 20, and 40 eV. Acoustic droplet ejection into 384-well plates enabled the scale-up, followed by standardized liquid chromatography–tandem mass spectrometry acquisition on quadrupole time-of-flight instruments. 

Beyond scale, the platform is designed to narrow the candidate lists that can follow exact-mass matching in large databases. Users can search by precursor mass, formula, name, structure, or fragment ions, then refine annotations through MS/MS matching, neutral-loss searches, in-source fragment filtering, and biological-context prioritization. A METLIN Core subset of around 31,000 frequently queried compounds is also included for faster routine searches. 

The resource also extends into targeted method development through METLIN-MRM, which derives multiple reaction monitoring transitions from empirical fragmentation behavior rather than rule-based or in silico predictions. Product-ion intensities across collision energies are used to nominate quantifier and qualifier ions, supporting quantitative assay development across chemically diverse small molecules. 

For metabolomics and related workflows, the update addresses a familiar bottleneck by pairing broad chemical coverage with standardized empirical spectra for more reliable small-molecule annotation. 

Storing Data in Protein 

A proof-of-concept system uses engineered proteins and LC–MS/MS sequencing to write and recover digital data 

Researchers have built a proof-of-concept protein data storage system that encodes digital information in engineered amino acid sequences and retrieves it through LC–MS/MS sequencing. 

The study addresses two challenges in using proteins as data carriers: how to express highly variable artificial sequences while preserving stability and solubility, and how to recover the complete sequence accurately enough to reconstruct the stored file. Rather than joining data-bearing peptides directly, the team embedded encoded amino acid sequences into a collagen-like protein template designed to improve expression and durability. 

For retrieval, the proteins were digested with trypsin and analyzed by LC–MS/MS. Custom software assigned peptide sequences from the tandem mass spectra, rebuilt the full protein sequence, and converted it back into binary data, with error correction used to recover missed or incorrectly read segments. 

The team first stored and recovered short text messages, then extended the approach to mixtures of multiple proteins encoding larger datasets. The collagen-like template improved expression compared with simpler peptide-fusion designs, and one data-bearing protein remained readable after high-temperature and strongly acidic treatments that rapidly degraded corresponding data-bearing DNA.  

The researchers also demonstrated protein-specific functions for storage and retrieval. Affinity tags allowed selected data-bearing proteins to be purified for random access, while a related tag-and-antibody strategy protected secret messages unless the correct molecular “key” was used. 

The approach remains at an early stage, but the study shows how protein engineering and LC–MS/MS sequencing could be combined to store and retrieve digital information in an alternative molecular format. “Moving forward, we aim to achieve mass storage capabilities, faster data writing and reading speeds, and further reductions in protein production costs,” said corresponding author Zhongping Yao in the team’s press release. 

Why p53 Needs the Full Site 

Native mass spectrometry shows that p53 needs both decameric half-sites of a 20-bp response element to form a stable tetrameric complex 

The tumor suppressor p53 requires the complete architecture of a 20-base-pair DNA response element, not just a single half-site, to form a stable tetrameric complex, according to a recent native MS study. 

P53 normally binds DNA as a tetramer, with two dimers engaging the two half-sites of a response element. To test how much of that architecture is required, researchers compared full-length wild-type p53 with p53L344A, a dimer-biased variant, across 37 DNA constructs. These included full p21 response elements, isolated half-sites, sequences with endogenous or random flanking extensions, and random DNA controls. 

Native mass spectrometry allowed the team to track which p53-DNA assemblies formed without disrupting the underlying noncovalent interactions. Complete 20-base-pair response elements supported tetrameric p53-DNA assembly, consistent with two p53 dimers assembling across the full site. Shorter or less specific constructs, including isolated half-sites, extended half-sites, and random 20-base-pair sequences, only supported dimeric binding. 

The same pattern held for wild-type p53 and the dimeric variant, suggesting that the flanking regions around the response element have little influence on initial complex formation. Instead, the decisive feature was the presence of both decameric half-sites, which allowed dimer-dimer interactions to stabilize the tetrameric p53 assembly. 

The findings point to dimer-dimer interactions as a key stabilizing step in p53-DNA recognition, helping explain why the full two-half-site response element is needed. 

A New Way to Split Overlapping Fingermarks 

The DESI-MSI workflow treats superimposed latent fingermarks as additive chemical maps rather than relying on optical ridge separation alone 

A new analysis of overlapping fingermarks shows that donor-specific chemical profiles can be recovered even when visual ridge patterns are superimposed. 

The study used desorption electrospray ionization mass spectrometry imaging (DESI-MSI) data from latent fingermarks lifted onto forensic gelatin supports. Each dataset contained up to about 200,000 pixels, with each pixel carrying intensity measurements across thousands of mass-to-charge channels. Rather than relying on optical separation of superimposed ridge patterns, the researchers treated the imaging data as a high-dimensional chemical map. 

To separate the overlapping deposits, the team applied non-negative matrix factorization (NMF) – an unsupervised chemometric method that decomposes the data into spatial components and their associated molecular signatures. Because mass spectrometry intensities are non-negative and overlapping fingermarks behave like additive chemical deposits, the authors argue that NMF is better matched to the problem than principal component analysis, which can mix spatial patterns through positive and negative loadings. 

Across five independently acquired two-donor samples, NMF produced spatially coherent components that localized to individual fingermarks, with clearer donor-specific boundaries than PCA. The associated molecular signatures could then be inspected for m/z features linked to each component, allowing separated ridge patterns and donor-associated chemical profiles to be assessed together. 

The authors note that real casework will still require validation across more varied surfaces, degraded traces, contaminants, and unknown contributor numbers. Nonetheless, the study gives forensic analysts a more automated route for separating chemically distinct fingermarks when visual inspection alone falls short. 

Newsletters

Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.

I have read and understand the Privacy Notice*

Subscribe