A recent paper in Nature, from the Church Lab at Harvard Medical School & the Wyss Institute introduces AGENTEX (automated genetic tRNA expansion), a robotic, multiplexed tool for building and testing new genetic codes in cell-free translation systems.

Radford et al. revisited a long-standing assumption that the tRNA 3′ CCA end, where amino acids attach and tRNAs dock into the ribosome, must be intact for aminoacylation. Using a new sequencing method they developed, tSCAN, the team varied the tRNA 3′ end across the entire E. coli tRNA set and found that most non-CCA tRNAs ("otRNAs") were still efficiently charged by natural synthetases, far more tolerant to mutation than previously believed.

Building on this, the authors engineered ribosomes (G2251C/G2553C, "CGA ribosomes") that specifically accommodate otRNAs while ignoring native tRNAs. Pairing otRNA pools with these orthogonal ribosomes gave rise to AGENTEX, run end-to-end on an automated OT-2 robotic platform. Using it, the team compressed the 64-codon code down to as few as 21–22 codons, incorporated a non-standard amino acid via amber suppression, and showed the compressed and standard genetic codes could run in parallel, without "crosstalk". Confirming this "zero crosstalk" required proteomics sensitive enough to catch rare mistranslation events. The team relied on PEAKS Studio 13 to search high-sensitivity LC-MS/MS data against custom libraries covering every possible amino acid substitution their engineered codes could introduce, showing that the ribosome, not the synthetase, is the real gatekeeper of genetic code fidelity.

Congratulations to the Church Lab on this remarkable accomplishment! This breakthrough technology overturns decades of understanding of how transfer RNAs (tRNAs) help turn the genetic code into proteins and takes a leap forward by making 34 codons customizable, allowing researchers to make proteins with up to 34 different amino acids much faster and safer.

BSI is delighted to witness the application of PEAKS software in this pioneering research within synthetic biology! We sincerely appreciate our ongoing collaboration with Dr. Bogdan Budnik and look forward to future advancements.

Read the full article here: Radford, F., Sapers, N., Burgess, H. M., Ort, L., Budnik, B., & Church, G. M. (2026) Automated prototyping of genetic codes. Nature. https://doi.org/10.1038/s41586-026-10949-y

A recent paper published in Nucleic Acids Research, from Western University in collaboration with BSI, highlights that naturally occurring human tRNA variants can drive systematic mistranslation in healthy cells, at levels that are high enough to matter for how we interpret proteomic data.

Rozik et al. measure mistranslation levels caused by a naturally occurring human tRNA anticodon variant, present in about 2% of the population, that introduces serine incorporation at phenylalanine codons. To do this, the authors introduced a clever dual GFP–mCherry reporter that converts translational errors into a quantitative fluorescent signal in live cells. By coupling GFP (as a production control) with a mistranslation-sensitive mCherry variant, amino acid mis-incorporation events in living human and mouse cells were detected and quantified. Using this reporter, the team showed that a single-nucleotide change (G35A) in a serine tRNA anticodon causes serine to be incorporated at phenylalanine codons—a clear violation of the canonical genetic code.

Importantly, this effect was confirmed using mass spectrometry, validating that the observed fluorescence directly reflects mistranslated protein. To support their work, the study also examined β-lymphocyte cell lines from the 1000 Genomes Project, enabling them to sample genetic diversity across the human population. Cells carrying the G35A allele not only expressed the mutant tRNA but consistently showed elevated mistranslation signals using the live-cell reporter—something not observed in cells with the wild-type tRNA.

These findings challenge the long-held assumption that healthy human cells strictly adhere to an error-free genetic code. Instead, they suggest that mistranslation is tolerated in humans, reshaping how we think about translation fidelity and protein quality control. BSI is proud to support and collaborate with academic and research partners on work that advances our understanding of human genetic and proteomic diversity and its functional impact.

Read the full article here: Rozik, P., Moore, H., Lant, J. T., Hoffman, K. S., Schultz, S. K., Afzal, B., Chan, P. P., Flynn, L. E., Heinemann, I. U., Lowe, T. M., O’Donoghue, P. (2026) Mistranslation from an endogenous tRNA variant in human pan-genome cell lines. Nucleic Acids Res. 54(5). doi:10.1093/nar/gkag224

Advancing de novo Antibody Sequencing with Hyperthermal Proteases and Next‑Gen Fragmentation

A recent Cell Systems study coming from our collaborators at CinderBio, led by Dr. Steven Yannone, in a joint work with Dr. Albert Heck's lab, highlights how highly efficient hyperthermal acidic proteases, combined with hybrid-fragmentation schemes, greatly boost confidence in de novo sequencing. Combining the HTA-based hyperthermal single-protease approach with powerful ZenoTOF EAciD fragmentation provides information-rich spectra and enables confident de novo antibody sequencing based on a single LC-MS run.

Conventional tryptic digestion often produces short peptides and limited coverage of complementarity‑determining regions (CDRs). As a result, de novo antibody sequencing typically requires multiple proteases and repeated MS analyses, increasing complexity and cost. In this study, hyperthermoacidic archaeal proteases were used to generate long, overlapping peptides ideally suited for PEAKS de novo analysis in a single digestion step.

We are excited to see PEAKS DeepNovo peptide sequencing algorithm efficiently handling the rich fragment ion series from EAciD to delivered high‑confidence peptide sequences and strong residue‑level support across antibody variable regions. Using confident peptide sequences from PEAKS Studio, full antibody sequences—including all CDRs—were reconstructed with higher confidence and fewer errors than conventional workflows.

BSI is proud to support these breakthroughs within our software portfolio, including new support for HTA‑proteases in our automated protein and antibody sequencing tool, PEAKS AB, enabling researchers to confidently sequence peptides from spectra generated by next‑generation fragmentation technologies and extract deeper insights from MS data.

If you’re interested in seeing how this fascinating integration can boost your research, reach out to us today!

Read the full article here: Pañeda, L. P., Kadavá, T., Shamorkina, T. M., Schulte, D., Pribil, P., Heidelberger, S., Narlock-Brand, A. M., Yannone, S. M., Snijder, J., & Heck, A. J. R. (2026). Deep coverage and extended sequence reads obtained with a single archaeal protease expedite de novo protein sequencing by mass spectrometry. Cell Systems. 17, 101536. doi:10.1016/j.cels.2026.101536

PEAKS Studio 13.1 shows off its application prowess in working with the new ZenoTOF 8600!

We are thrilled to see PEAKS Studio 13.1 be featured in the paper “Scanning DIA on the ZenoTOF 8600 system enables ultra-sensitive and quantitative proteomics from single cells to post-translational modifications in a compact platform” and used to benchmark the proteomics performance of the state-of-the-art platform.

In this paper, Heymann et al., led by Dr. Matthias Mann, demonstrate the effectiveness of the ZenoTOF 8600 by benchmarking its performance across a broad range of proteomics applications, including  high-throughput and single-cell proteome coverage, quantitative accuracy, targeted sensitivity, and post‑translational modification analysis.

Using scanning quadrupole DIA (ZT Scan DIA), the authors show improved peptide and protein identifications and quantitative reproducibility compared with conventional DIA, while demonstrating robust performance when using PEAKS Studio 13.1 across applications ranging from ultra‑high‑throughput bulk proteomics and single‑cell analysis to low‑attomole targeted quantitation and disease‑relevant phosphorylation profiling.

The PEAKS team is honoured to be part of this rigorous evaluation of the ZenoTOF 8600, with PEAKS Studio 13.1 serving as a reliable benchmark for ultra-sensitive and quantitative proteomics. At BSI, we remain committed to advancing accurate, real‑time proteomics analysis through deep‑learning–powered solutions that enable researchers to fully realise the potential of next‑generation instrumentation.

Interested in exploring the extensive capabilities of PEAKS Studio? Reach out to us today!

Read the full article here: Heymann, T., Oliinyk, D., Henneberg, L., Lorenz, M.B., Eikmeier, N., Thielert, M., Oeller, M.,  Grauvogel, L., Sitron, C.S., Loyd, B., Le Blanc, Y., Bloomfield, N., Batruch, I., Causon, J., Chelur, A., Ivosev, G., Tran, K., Talamantes, T., Schneider, B., Castro-Perez, J., Mann, M. (2026). Scanning DIA on the ZenoTOF 8600 system enables ultra-sensitive and quantitative proteomics from single cells to post-translational modifications in a compact platform. bioRxiv. doi:10.64898/2026.03.12.711261

AI is reshaping how we characterise complex antibody therapeutics—and PolySeq.AI is part of that conversation!

We’re excited to see PolySeq.AI, BSI’s newest AI-driven solution, featured in the recent review paper, Polyclonal Antibody Therapeutics: Analytical Innovations and Regulatory Perspectives for Addressing Heterogeneity Challenges.

As Kumar et al. highlight, continued progress in the characterisation of complex biological products depends not only on advances in analytical workflows and instrumentation, but also on the ability to extract meaningful insight from the increasingly large and complex datasets they generate. There is a growing need for specialised algorithms and tools capable of handling the scale and heterogeneity of modern mass spectrometric data.

In particular, the review cites our work on PolySeq.AI, describing how it was applied to bottom‑up proteomic analyses for de novo sequencing of antibody mixtures. PolySeq.AI is an automated, AI‑driven platform that enables full de novo sequencing of complex polyclonal antibodies by integrating intact mass, middle‑down, and bottom‑up mass spectrometry—without relying on external databases. It achieves >99% sequencing accuracy, making antibody characterisation fast, reproducible, and practical at scale.

While positioned as a proof of concept, the authors emphasise that this study: “highlights the transformative potential of AI‑based tools to enable sequence‑level characterization of pAb products.” We’re proud to see PolySeq.AI recognised in this important discussion and to contribute to the evolving analytical toolbox needed to address the unique challenges of polyclonal antibody therapeutics.

Reach out to discover how our AI‑driven solutions can take your data analysis even further!

Read the review: Kumar, S., Tini, A., Tengattini, S., Rinaldi, F., Calleri, E., Massolini, G., & Temporini, C. (2026). Polyclonal Antibody Therapeutics: Analytical Innovations and Regulatory Perspectives for Addressing Heterogeneity Challenges. Analytical Chemistry. doi:10.1021/acs.analchem.5c06531

Read on PolySeq.AI: Sequencing of Polyclonal Antibodies by Integrating Intact Mass, Middle–Down, and De Novo Bottom–Up Mass Spectrometry

Cancer cells overexpressing CD47 receptors often evade innate immune responses as CD47 interacts with SIRPα, an inhibitory immunoreceptor expressed on phagocytes, protecting them from phagocytosis. Anti-CD47 antibodies can bind to these receptors and block this signal, enabling macrophages to target and destroy cancer cells. While promising, clinical benefits have been limited, highlighting the need for deeper understanding of their mechanisms.

A key question is whether anti-CD47 antibodies act solely by blocking CD47/SIRPα via the Fab domain or also through Fc–FcγR interactions. Osorio et al. explored this by engineering antibodies with Fc regions of varying FcγR affinities. Antibodies with higher FcγR affinity demonstrated superior antitumour activity, markedly reducing primary tumour growth and lung metastases across multiple mouse models. These findings highlight the critical role of Fc-mediated immune engagement in enhancing therapeutic efficacy.

BSI was proud to contribute our antibody sequencing services, DeepAB, to perform high-accuracy mass spectrometry-based de novo sequencing of the anti-CD47 variable regions, which were then used to engineer antibodies with modified Fc regions. Our service provided in-depth sequence coverage using multi-enzyme digests and high-resolution mass spectrometry data, analysed in our PEAKS AB software. The sequences were validated by matching the theoretical mass of bottom-up sequence assemblies to intact mass measurements. In addition, leucine/isoleucine differentiation, PTM analysis, and glycan profiling are including in this service.  Taken together, our DeepAB workflow ensures highly accurate de novo sequencing and in-depth antibody characterisation.

We are honored to be part of such impactful research and proud to support innovations that advance cancer immunotherapy. If you’re interested in our antibody sequencing services or want to see how we can support your research, contact us today to learn how BSI can help accelerate your discoveries!

Read the full article here: Osorio, J. C., Smith, P., Knorr, D. A., & Ravetch, J. V. (2023). The antitumor activities of anti-CD47 antibodies require Fc-FcγR interactions. Cancer Cell, 41(12). doi:10.1016/j.ccell.2023.10.007

Proteins undergo post-translational modifications (PTMs) such as phosphorylation, ubiquitination, acetylation, and methylation. These modifications play essential roles for numerous biological functions, including signalling, regulation, protein activity, and gene expression control. The pathological development of many neurological diseases is known to be influenced by PTM dysregulation. Accurate identification and characterisation of these modifications are critical for a comprehensive understanding of cellular biology and the mechanisms underlying these diseases. A recent study in Biomeditsinskaya Khimiya (Biomedical Chemistry) dives deeper into the involvements of PTMs in the pathogenesis of Temporal Lobe Epilepsy (TLE) in adults.

To achieve this, Miroshnichenko et al. harnessed the power of PEAKS Studio to conduct a comparative proteome analysis of hippocampal tissue from patients with sclerotic and non-sclerotic temporal lobe epilepsy, as well as nonepileptic controls, to uncover differences in protein PTM levels. Thanks to PEAKS advanced PTM algorithm, which takes advantage of de novo partial tags, researchers were able to specify physiological PTMs, such as methylation of KRH, acetylation of K, acetylation of Protein N-terminal peptide, and citrullination of N in addition to more common oxidation of M and STY phosphorylation, without a significant effect on the search space. This allowed them to uncover deeper biological insight into the mechanism of pathological change in TLE. Scientists identified 53 proteins with significant changes in the levels of post-translational modifications linked to neurological disease pathways and epileptogenesis. By identifying site-specific PTMs, this study highlights molecular alterations that could serve as potential biomarkers for disease progression.

Read the full article here: Miroshnichenko, Yu. V., Rybina, A. V., Skvortsov, V. S. (2025). Identification of proteins with variable levels of post-translational modifications in human temporal lobe epilepsy. Biomeditsinskaya Khimiya, 71(5), 351-363. doi:10.18097/PBMCR1612

A new study in Cell Reports Methods by Cell Press takes immunopeptidomics in an exciting new direction. While traditional approaches focus on peptides presented by classical MHC molecules, the researchers asked a different question: how can we profile non-peptidic antigens?

This study expands the field by profiling small-molecule ligands bound to MR1, a non-classical MHC class I–like molecule. Using a clever cross-linking approach, the team covalently stabilised the fleeting interaction between MR1 and vitamin B6–derived metabolites, making these complexes detectable by mass spectrometry.

They then harnessed 𝗣𝗘𝗔𝗞𝗦 𝗦𝘁𝘂𝗱𝗶𝗼 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 to identify these small molecules as post-translational modifications on a reporter peptide, pinpointing their mass, composition, and specificity. This workflow opens new possibilities for studying unconventional antigen presentation and T-cell recognition.

Read the full article here: Schmidlin, T., Behiry, E., Thomas, H., Dolton, G., Marino, F., Hasan, S., ... & Ternette, N. (2025). MR1-ligand cross-linking identifies vitamin B6 metabolites as TCR-reactive antigens. Cell Reports Methods. doi:10.1016/j.crmeth.2025.101120

A recent publication in Science has identified the Harbin cranium, a remarkably well-preserved fossil from northeastern China, as belonging to a Denisovan individual. Dated to at least 146,000 years ago, this cranium represents the most complete Denisovan-associated specimen analysed to date, offering new insight into the physical characteristics of this ancient human group.

Because ancient DNA was not preserved in the specimen, the researchers turned to paleoproteomics to investigate its identity. Proteins were extracted from the petrous portion of the skull and analysed using mass spectrometry. The data were processed with PEAKS Online, which enabled the identification of over 20,000 peptides and 95 endogenous proteins. Several of these proteins contained amino acid variants that matched those previously found in Denisovan genomes, allowing the team to make a confident taxonomic assignment.

This study marks the first time Denisovan molecular markers have been directly linked to a complete cranium, bridging the gap between genetic data and skeletal morphology. It also demonstrates the value of proteomic analysis in cases where ancient DNA is not recoverable, and reinforces the growing role of protein-based methods in human evolutionary research.

Read the full article in Science: https://doi.org/10.1126/science.adu9677

Tian, H., Li, G., Chiu, C. K., Li, E., Chen, Y., Zhu, T., ... & Kang, X. (2025). Rapid and direct discovery of functional tumor specific neoantigens by high resolution mass spectrometry and novel algorithm prediction. Cell Insight, 100251. doi:10.1016/j.cellin.2025.100251

Accelerating Personalized Cancer Vaccines with LC-MS Immunopeptidomics and PEAKS

Personalized cancer vaccines are transforming oncology by targeting tumor-specific neoantigens - peptides uniquely presented on the surface of cancer cells. Identifying these neoantigens rapidly and accurately is a critical step in designing effective, patient-specific immunotherapies.

While next-generation sequencing (NGS) plays a foundational role in cancer vaccine development, it predicts neoantigen candidates based on DNA or RNA sequences, not the actual peptides displayed on the cell surface. In contrast, LC-MS–based immunopeptidomics directly detects peptides presented by MHC complexes, making it a more reliable approach for identifying functional neoantigens.

A Six-Week Workflow from Discovery to Validation

In a recent study, researchers implemented a streamlined six-week workflow, from tumor tissue to validated neoantigen candidates, to accelerate personalized cancer vaccine development. The key to this rapid turnaround was integrating RNA-seq and LC-MS data from the same tumor sample, providing both transcriptomic and immunopeptidomic insights.

PEAKS Online played a central role in this integration, enabling scientists to search LC-MS immunopeptidome data against both:

• Public protein databases (UniProt), and
• Personalized, patient-specific protein sequences derived from RNA-seq.

The Power of De Novo Sequencing

In addition to traditional database searching, the team leveraged PEAKS' de novo sequencing capabilities to uncover peptides not found in standard databases. Using database-identified peptides as ground truth, they estimated that approximately 60% of de novo peptides were correctly sequenced, highlighting their potential to reveal novel, patient-specific antigens missed by conventional search methods.

To prioritize candidates for further testing, the team used two different algorithms to predict peptide-MHC binding affinities and selected top-ranked peptides for in vitro validation.

Why In Vitro Validation Still Matters

Interestingly, in vitro T-cell response assays revealed that the top five predicted candidates did not trigger a strong immune response, while lower-ranked peptides, candidates 6 through 15, did. This result underscores a crucial lesson in neoantigen discovery: computational predictions must be validated experimentally. Immunogenicity cannot always be inferred from binding scores alone.

Conclusion

This study reinforces the value of LC-MS–based immunopeptidomics, particularly when paired with advanced software like PEAKS for database and de novo sequencing. The ability to discover and validate tumor-specific immunopeptides in just six weeks is a promising step forward in making personalized cancer vaccines faster, more accurate, and more accessible.