Biology subjects

Valli, A. A.

Publications and source records attributed to Valli, A. A..

7 recordsLinked to original sources

bioRxiv2026

On the robustness of scRNA-seq foundation models for plant perturbation response prediction under cross-experiment shift

Foundation models for single-cell transcriptomics promise to learn generalizable representations of cellular states. However, recent evidence suggests they often fail to outperform simple machine learning baselines. Furthermore, their ability to generalize across unseen experimental conditions remains poorly understood, particularly in plants, where rigorous evaluation beyond cell type annotation and batch integration is lacking. To address this, we introduce an Arabidopsis thaliana foundation model, scAraFM, and benchmark it across several perturbation conditions under three increasingly challenging protocols: random splits from a single experiment, replicate-based splits, and cross-experiment transfer learning. We found that random splits overestimate performance by up to 30 points relative to cross-experiment evaluations. Across representation strategies, preserving gene identity consistently outperforms the standard pooled embeddings. Moreover, simple baselines using raw reads remain competitive in single-experiment settings, challenging current claims of universal advantage of foundation models. In contrast, under cross-experiment transfer, pretrained representations show added value, particularly with few labelled samples, suggesting that the benefits of foundation models emerge precisely in the regimes that matter for practical deployment. Overall, our results demonstrate that conclusions about foundation models depend critically on the evaluation design, and that preserving per-gene structure aids generalization in downstream tasks, supporting robust predictions across unseen experimental contexts.

bioinformatics↗

bioRxiv2026

Discovery of novel members of the Potyviridae family reveals expanded diversity, a broad host range, and evidence of fungal and oomycete infections

The family Potyviridae represents the largest and most economically important group of plant-infecting RNA viruses. Despite extensive study of crop-associated members, the full diversity, host range, and evolutionary history of potyvirids remain poorly understood. Here, we conducted a large-scale mining of publicly available RNA-seq datasets to systematically search for novel potyvirid sequences. This approach enabled the identification and assembly of 47 previously undescribed members of the family, distributed across eight recognized genera and, importantly, two putative new genera. Beyond expanding the known genetic diversity of Potyviridae, our analyses revealed a distinct and deeply divergent lineage of potyvirid-like viruses associated with fungi and oomycetes, for which we propose the genus Macrophovirus. These viruses possess compact genomes and atypical genomic organizations, including the absence of canonical plant cell-to-cell movement factors and the presence of HCPro-like proteins arranged in tandem. Comparative structural and phylogenetic analyses indicate that these leader proteases are more closely related to fungal hypoviral counterparts than to canonical potyvirid HCPro factors. Together, our findings substantially expand the host range of Potyviridae, provide compelling evidence that potyvirid-like viruses likely infect fungi and oomycetes in nature, and offer new insights into the evolutionary pathways that have shaped this major virus family.

evolutionary biology↗

bioRxiv2025

Turnip mosaic virus drives selective filtering and community reassembly in the Arabidopsis thaliana root microbiome in a genotype-specific manner

Plant root microbiomes play a central role in plant health, yet their responses to viral infection remain poorly understood. Here, we investigated how Turnip mosaic virus (TuMV) alters root-associated bacterial and fungal communities in two Arabidopsis thaliana genotypes (Col-0 and Mar-12) grown in natural soils. Using 16S and ITS amplicon sequencing, we assessed changes in diversity, taxonomic composition, enriched microbial taxa, and co-occurrence network structure to distinguish between plant-mediated recruitment ("cry-for-help") and pathogen-induced dysbiosis. TuMV infection caused a pronounced reduction in bacterial diversity and a restructuring of bacterial community composition, whereas fungal communities remained largely stable. Viral infection also led to genotype-specific shifts in enriched bacterial genera, with opportunistic and stress-tolerant taxa proliferating differently in each genotype. Despite the initial perturbation, bacterial networks recovered connectivity and, in some cases, reached higher complexity than those of healthy plants, indicating strong microbial resilience. Together, these results reveal that TuMV infection acts as a selective filter on bacterial, but not fungal, root communities and that the surviving taxa can reorganize into functional networks. Our study provides one of the most comprehensive assessments of virus-induced microbiome restructuring and highlights the importance of host genotype in shaping microbial responses to biotic stress.

plant biology↗

bioRxiv2025

Natural self-attenuation of pathogenic viruses by deleting the silencing suppressor coding sequence for long-term plant-virus coexistence

Potyviridae is the largest family of plant-infecting RNA viruses. All members of the family (potyvirids) have positive-sense single-stranded RNA genomes, with polyprotein processing as the main expression strategy. The 5-proximal regions of their genomes encode two types of leader proteases: the serine protease P1 and the cysteine protease HCPro. However, their arrangement and sequence composition vary greatly among genera or even species. The leader proteases play multiple important roles in different potyvirid-host combinations, including RNA silencing suppression and virus transmission. Here, we report that viruses in the genus Arepavirus, which encode two HCPro leader proteases in tandem (HCPro1-HCPro2), can naturally lose the coding sequences for these two proteins during infection. Notably, this loss is associated with a shift in foliage symptoms from severe necrosis to mild chlorosis or even asymptomatic infections. Further analysis revealed that the deleted region is flanked by two short repeated sequences in the parental isolates, suggesting that recombination during virus replication likely drives this genomic deletion. Reverse genetic approaches confirmed that the loss of leader proteases weakens RNA silencing suppression and other critical functions. A field survey of areca palm trees displaying varied symptom severity identified a transitional stage in which full-length viruses and deletion mutants coexist in the same tree. Based on these findings, we propose a scenario in which full-length isolates drive robust infections and facilitate plant-to-plant transmission, eventually giving rise to leader protease-less variants that mitigate excessive damage to host trees, allowing long-term coexistence with the perennial host. To our knowledge, this is the first report of potyvirid self-attenuation via coding sequence loss. Author summaryPlant viruses typically persist throughout the lifespan of their host plants, employing multiple strategies to ensure long-term survival. This study reveals an unusual self-attenuation mechanism in which potyvirids, likely through recombination, discard a large genomic fragment containing the RNA silencing suppressor coding sequence. This deletion reduces viral pathogenicity, enabling a peaceful long-term virus-plant coexistence. Meanwhile, the attenuated viruses may function as natural vaccines, protecting plants from reinfection by the full-length pathogenic strains or related variants. These findings highlight a direct link between viral evolution and long-term coexistence, suggesting that such adaptations may promote mutual benefits. Understanding these dynamics could provide valuable insights into virus-host interactions and inspire new approaches to plant disease management.

pathology↗

bioRxiv2024

Reconceptualizing programmed transcriptional slippage in RNA viruses

RNA viruses have evolved sophisticated strategies to exploit the limited encoded information within their typically compact genomes. One of such, named programmed transcriptional slippage (PTS), is defined by the insertion of an additional A at An motifs (n [≥] 6) of newly synthetized viral transcripts to get access to overlapping open reading frames (ORFs). Although key proteins from Ebolavirus and potyvirids (members of the Potyviridae family) are expressed via PTS, available information about this phenomenon is very scarce. Here, by using diverse experimental approaches and a collection of plant/virus combinations, we discover cases in which PTS does not fit with its current definition. In summary, we observe (i) high rate of single nucleotide deletions at slippage motifs, (ii) overlapping ORFs acceded by slippage at an U8 stretch, and (iii) significant changes in slippage rates induced by factors not related to cognate viruses. Moreover, a survey of full-genome sequences from potyvirids shows a widespread occurrence of species-specific An/Un (n [≥] 6) motifs. Even though many of them, but not all, lead to the production of truncated proteins rather than access to overlapping ORFs, these results suggest that slippage motifs appear more frequently than expected and play relevant roles during virus evolution. In conclusion, our data prompt to broaden PTS definition in RNA viruses. Considering the potential of this phenomenon to expand the viral proteome by acceding to overlapping ORFs and/or producing truncated proteins, a revaluation of PTS significance during infections of RNA viruses is required. IMPORTANCEProgrammed transcriptional slippage (PTS) is used by RNA viruses as another strategy to maximise the coding information in their genomes. This phenomenon is based on a peculiar feature of viral replicases: they insert an untemplated A in An motifs (n [≥] 6) in a small fraction of newly synthesised viral RNAs. As a consequence, ribosomes can get access to overlapping open reading frames (ORFs) when translating those particular transcripts. Here, using plant-infecting RNA viruses as models, we discover cases challenging the previously stablished definition of viral PTS, prompting us to reconsider and redefine this expression strategy. An interesting conclusion from our study is that PTS might be more relevant during RNA virus evolution and infection processes than previously assumed.

molecular biology↗

bioRxiv2021

Maf/ham1-like pyrophosphatases, host-specific partners of viral RNA-dependent RNA polymerases

Cassava brown streak disease (CBSD), dubbed the "Ebola of plants", is a serious threat to food security in Africa caused by two viruses of the family Potyviridae: cassava brown streak virus (CBSV) and Ugandan (U)CBSV. Intriguingly, U/CBSV, along with another member of this family and one secoviridae, are the only known RNA viruses encoding a protein of the Maf/ham1-like family, a group of widespread pyrophosphatase of non-canonical nucleotides (ITPase) expressed by all living organisms. Despite the socio-economic impact of CDSD, the relevance and role of this atypical viral factor has not been yet established. Here, using an infectious cDNA clone and reverse genetics, we demonstrate that UCBSV requires the ITPase activity for infectivity in cassava, but not in the model plant Nicotiana benthamiana. HPLC MS/MS experiments showed that, quite likely, this host-specific constraint is due to an unexpected high concentration of non-canonical nucleotides in cassava. Finally, protein analyses and experimental evolution of mutant viruses indicated that keeping a fraction of the yielded UCBSV ITPase covalently bound to the viral RNA-dependent RNA polymerase (RdRP) optimizes viral fitness, and this seems to be a feature shared by the other members of the Potyviridae family expressing Maf/ham1-like proteins. All in all, our work (i) reveals that the over-accumulation of non-canonical nucleotides in the host might have a key role in antiviral defense, and (ii) provides the first example of an RdRP-ITPase partnership, reinforcing the idea that RNA viruses are incredibly versatile at adaptation to different host setups.

molecular biology↗

bioRxiv2020

Mapping the Algal Secret Genome: The small RNA Locus Map for Chlamydomonas reinhardtii

Small (s)RNAs play crucial roles in the regulation of gene expression and genome stability across eukaryotes where they direct epigenetic modifications, post-transcriptional gene silencing, and defense against both endogenous and exogenous viruses. The green alga Chlamydomonas reinhardtii is a well-studied unicellular alga species with sRNA-based mechanisms that are distinct from those of land plants. It is, therefore, a good model to study sRNA evolution but a systematic classification of sRNA mechanisms is lacking in this and any other algae. Here, using data-driven machine learning approaches including Multiple Correspondence Analysis (MCA) and clustering, we have generated a comprehensively annotated and classified sRNA locus map for C. reinhardtii. This map shows some common characteristics with higher plants and animals, but it also reveals distinct features. These results are consistent with the idea that there was diversification in sRNA mechanisms after the evolutionary divergence of algae from higher plant lineages.

bioinformatics↗