Biology subjects

Mann, B.

Publications and source records attributed to Mann, B..

4 recordsLinked to original sources

bioRxiv2025

seqLens: optimizing language models for genomic predictions

Understanding evolutionary variation in genomic sequences through the lens of language modeling has the potential to revolutionize biological research. Yet to maximize the utility of language modeling in genomics, we must overcome computational challenges in tokenization and model architecture adapted to diverse genomic features across evolutionary timescales. In this study, we investigated key elements in genomic language modeling (gLM), including tokenization, pretraining datasets, fine-tuning approaches, pooling methods, and domain adaptation, and applied the language models to diverse genomic data. We gathered two evolutionarily distinct pretraining datasets: one consisting of 19,551 reference genomes, including over 18,000 prokaryotic genomes (115B nucleotides) and the remainder eukaryotic genomes, and another more balanced dataset with 1,354 genomes, including 1,166 prokaryotic and 188 eukaryotic reference genomes (180B nucleotides). We trained five byte-pair encoding tokenizers and pretrained 52 gLMs, systematically comparing different architectures, hyperparameters, and classification heads. We introduce seqLens, a family of models based on disentangled attention with relative positional encoding, which outperforms relatively similar-sized models in 13 of 19 benchmarking phenotypic predictions. We further explore continual pretraining, domain adaptation, and parameter-efficient fine-tuning methods to assess trade-offs between computational efficiency and accuracy. Our findings demonstrate that relevant pretraining data significantly boost performance, alternative pooling techniques can enhance classification, tokenizers with larger vocabulary sizes negatively impact generalization, and gLMs are capable of understanding evolutionary relationships. These insights provide a foundation for optimizing genomic language models for identifying diverse evolutionary genomic features and improving genome annotations.

bioinformatics↗

bioRxiv2023

Imprinted anti-hemagglutinin and anti-neuraminidase antibody responses after childhood infections of A(H1N1) and A(H1N1)pdm09 influenza viruses

Immune imprinting is a driver known to shape the anti-hemagglutinin (HA) antibody landscape of individuals born within the same birth cohort. With the HA and neuraminidase (NA) proteins evolving at different rates under immune selection pressures, anti-HA and anti-NA antibody responses since childhood influenza infections have not been evaluated in parallel at the individual level. This is partly due to the limited knowledge of changes in NA antigenicity, as seasonal influenza vaccines have focused on generating neutralising anti-HA antibodies against HA antigenic variants. Here we systematically characterised the NA antigenic variants of seasonal A(H1N1) viruses from 1977 to 1991 and completed the antigenic profile of N1 NAs from 1977 to 2015. We identified that NA proteins of A/USSR/90/77, A/Singapore/06/86, and A/Texas/36/91 were antigenically distinct and mapped N386K as a key determinant of the NA antigenic change from A/USSR/90/77 to A/Singapore/06/86. With comprehensive panels of HA and NA antigenic variants of A(H1N1) and A(H1N1)pdm09 viruses, we determined hemagglutinin inhibition (HI) and neuraminidase inhibition (NI) antibodies from 130 subjects born between 1950-2015. Age-dependent imprinting was observed for both anti-HA and anti-NA antibodies, with the peak HI and NI titers predominantly detected from subjects at 4-12 years old during the year of initial virus isolation, except the age-independent anti-HA antibody response against A(H1N1)pdm09 viruses. More participants possessed antibodies that reacted to multiple antigenically distinct NA proteins than those with antibodies that reacted to multiple antigenically distinct HA proteins. Our results support the need to include NA proteins in seasonal influenza vaccine preparations. IMPORTANCESeasonal influenza vaccines have aimed to generate neutralizing anti-HA antibodies for protection since licensure. More recently, anti-NA antibodies have been established as an additional correlate of protection. While HA and NA antigenic changes occurred discordantly, the anti-HA and anti-NA antibody profiles have rarely been analysed in parallel at the individual level, due to the limited knowledge on NA antigenic changes. By characterizing NA antigenic changes of A(H1N1) viruses, we determined the anti-HA and anti-NA antibody landscape against antigenically distinct A(H1N1) and A(H1N1)pdm09 viruses using sera of 130 subjects born between 1950-2015. We observed age-dependent imprinting of both anti-HA and anti-NA antibodies against strains circulated during the first decade of life. 67.7% (88/130) and 90% (117/130) of participants developed cross-reactive antibodies to multiple HA and NA antigens at titers [≥]1:40. With slower NA antigenic changes and cross-reactive anti-NA antibody responses, including NA protein in influenza vaccine preparation may enhance vaccine efficacy. (150 words)

microbiology↗

bioRxiv2021

Fetal neural progenitors process TLR signals from bacterial components to enhance proliferation and rework brain development

Bacterial cell wall, a universal pathogen-associated molecular pattern (PAMP), crosses the placenta into the fetal brain. We determined that PAMPs interact with TLR2/6 on murine fetal neural progenitor cells (NPCs) to induce overexpansion of all neocortical layers leading to a larger, folded cortex and abnormal postnatal behavior. The NPC overexpansion originated at E10 and targeted ventricular radial glia (vRG), the primary NPC, by shortening cell cycle and increasing self-renewal. The mechanism involved two novel signaling pathways in NPCs mediated by recognition of bacterial PAMPs by TLR2/6 including: a) loss of primary cilia, activation of hedgehog signaling, and increased FOXG1 and b) increased PI3K/AKT activity. These findings reveal PAMP/TLR2/6 acts as a morphogen in fetal neurodevelopment. In addition, the loss of Tlr2 or Tlr6 without pathogenic challenge, increased the number of neurons, establishing the requirement for an endogenous TLR2 signal for normal neurodevelopment in the embryo.

neuroscience↗

bioRxiv2019

A fruitful endeavor: scent cues and echolocation behavior used by Carollia castanea to find fruit

Frugivores have evolved sensory and behavioral adaptations that allow them to find ripe fruit effectively, but the relative importance of different senses in varying foraging scenarios is poorly known. Within Neotropical ecosystems, short-tailed fruit bats (Carollia: Phyllostomidae) are abundant nocturnal frugivores, relying primarily on plants of the genus Piper as a food resource. Previous research has demonstrated Carollia employ olfaction and echolocation to locate Piper fruit, but it is unknown how their sensory use and foraging decisions are influenced by the complex diversity of chemical cues that fruiting plants produce. Using wild C. castanea and their preferred food, Piper scintillans, we conducted behavioral experiments to test two main hypotheses: (1) foraging decisions in C. castanea are primarily driven by ripe fruit scent and secondarily by vegetation scent, and (2) C. castanea re-weight their sensory inputs to account for available environmental cues, such that bats rely more heavily on echolocation in the absence of adequate scent cues. Our results suggest that C. castanea requires olfactory information and relies almost exclusively on ripe fruit scent to make foraging attempts. Ripe fruit scent is chemically distinct from vegetation scent in P. scintillans, with a greater abundance of {beta}-caryophyllene, germacrene D and {beta}-elemene, and a few unique compounds. Although variation in echolocation call parameters was independent of scent cue presence, bats emitted longer and more frequent echolocation calls in trials where no fruit scent was present. Altogether, these results highlight the adaptations, plasticity, and potential constraints in the sensory system of neotropical fruit bats. SIGNIFICANCE STATEMENTLittle is known about the relative importance of different senses and which plant cues are most important for fruit location by frugivores. We conducted behavioral experiments on short-tailed fruit bats (Carollia castanea), which use a combination of olfaction and echolocation to find ripe fruit, and their preferred food source (Piper scintillans) to test (1) which plant scent cues drive food selection and (2) if bats alter their echolocation behaviors based on which scent cues are present. We find that C. castanea rely almost exclusively on ripe fruit scent to forage, and echolocate more frequently when fruit scent is absent. Ripe fruit scent is chemically different from vegetation scent in P. scintillans, potentially providing a clear signal of food availability to mutualistic bats. These results highlight the sensory adaptations and behavioral flexibility of fruit bats as they navigate the cues provided by fruiting plants.

ecology↗