Biology subjects

Ronco, M.

Publications and source records attributed to Ronco, M..

3 recordsLinked to original sources

bioRxiv2026

Evo 2 as a classification machine: evidence from in-context learning and mechanistic interpretability

In-context learning (iCL) is an emergent capability of Large Language Models (LLMs), allowing them to perform new tasks at inference time using prompt-injected examples. While extensively studied in LLMs, the boundaries of its capabilities and underlying mechanisms remain poorly characterised in genomic Language Models (gLMs). Here, we map the operating regime of Evo 2, a nucleotide-level foundation gLM, across five binary classification tasks spanning biological and artificial sequences. We observe robust iCL on shorter natural sequences (F1=0.902 for miRNA, 0.785 for Toxins), degrading with sequence length, and collapsing at kilobase scale. Strikingly, we find no benefit in model scaling, as the 7B model systematically outperforms the 40B variant. We further find that perplexity, a widely used gLM performance proxy, poorly predicts accuracy. Mechanistic interpretability reconciles these observations: the logit-lens profiling suggests a prediction-generalisation trade-off, while Jacobian Scope indicates models might track prompts' structure rather than the signal-carrying content.

genomics↗

bioRxiv2026

Resolving power as a measure of protein stability prediction

Rank correlation and discriminative resolution are distinct properties of a stability predictor, yet current benchmarks measure only the first. Here we introduce resolving power: the {Delta}{Delta}G separation at which a predictor orders two variants correctly three times in four. Applied to BioEmu across twelve wild-type domains and three supervised regressors on an independent twenty-seven, resolving powers cluster near 1 kcal mol -1 --- comparable to the median variant-pair separation the benchmarks ask these models to rank, so most pairs cannot be ordered. A closed form ties this resolution to the calibration slope ; Spearman correlations account for only 31% of its variance. BioEmu extracts {Delta}G from the ratio of folded and unfolded populations, so what it samples of each becomes a question. Its folded populations reproduce experimental radii of gyration within half an [a]ngstrom, whereas those it labels unfolded are collapsed chains roughly 20 to 40% below a sequence-specific random-coil reference. The ensemble still carries variant differences on panels where the sampled unfolded fraction approaches zero, but the log-odds counting readout runs out of range. Scored by a learned energy, those same conformations recover resolution on ten of twelve panels, with the gain separated from noise on four, without retraining.

bioinformatics↗

bioRxiv2021

Integrative proteogenomics for differential expression and splicing variation in a DM1 mouse model

Dysregulated mRNA splicing is involved in the pathogenesis of many diseases including cancer, neurodegenerative diseases, and muscular dystrophies such as myotonic dystrophy type 1 (DM1). Comprehensive assessment of dysregulated splicing on the transcriptome and proteome level has been methodologically challenging, and thus investigations have often been targeting only few genes. Here, we performed a large-scale coordinated transcriptomic and proteomic analysis to characterize a DM1 mouse model (HSALR) in comparison to wild-type. Our integrative proteogenomics approach comprised gene- and splicing-level assessments for mRNAs and proteins. It recapitulated many known instances of aberrant mRNA splicing in DM1 and identified new ones. It enabled the design and targeting of splicing-specific peptides and confirmed the translation of known instances of aberrantly spliced disease-related genes (e.g. Atp2a1, Bin1, Ryr1), complemented by novel findings (e.g. Ywhae, Flnc, Svil). Comparative analysis of large-scale mRNA and protein expression data showed quantitative agreement of differentially expressed genes and splicing patterns between disease and wild-type. We hence propose this work as a suitable blueprint for a robust and scalable integrative proteogenomic strategy geared towards advancing our understanding of splicing-based disorders. With such a strategy, splicing-based biomarker candidates emerge as an attractive and accessible option, as they can be efficiently asserted on the mRNA and protein level in coordinated fashion. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=150 SRC="FIGDIR/small/443842v2_ufig1.gif" ALT="Figure 1"> View larger version (26K): org.highwire.dtl.DTLVardef@593246org.highwire.dtl.DTLVardef@1f3b5bforg.highwire.dtl.DTLVardef@12113b3org.highwire.dtl.DTLVardef@947d8f_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗