Biology subjects

von der Haar, T.

Publications and source records attributed to von der Haar, T..

8 recordsLinked to original sources

bioRxiv2026

A Universal, AI-based Design Framework for Efficient Manufacturing of mRNA Therapeutics

The growth of mRNA therapeutics is limited by bespoke manufacturing processes. To overcome this barrier to access and innovation, we introduce an AI-driven framework that decouples sequence design from manufacturing, analogous to the universal design principles that revolutionized the semiconductor industry. We performed a large-scale screen, quantifying the in vitro transcription (IVT) efficiency of one million diverse sequences. We then trained an interpretable deep learning model that accurately predicts manufacturability from sequences alone and learned underlying molecular mechanisms. An algorithm using this model prospectively improved the IVT yield of a vaccine and a gene-editing therapeutic by over 7.5-fold. Co-optimization for manufacturability and translation efficiency provided improvements over state-of-the-art commercial mRNA products. Our AI-driven framework establishes a universal design paradigm with the promise to democratize and accelerate the development of mRNA medicines, potentially unlocking a new era in biotechnology.

bioengineering↗

bioRxiv2025

Engineered yeast biofilms deliver plastic degrading enzymes to PET substrates

Enzymes that degrade PET and other plastics are being discovered with increasing pace, and engineering approaches further increase their activity. Most proposed applications for such enzymes require their purification, enabling the regeneration of monomeric subunits suitable for the production of new plastic. However, large amounts of plastics exist as environmental contaminants, and are unlikely to be suitable for standard recycling. Such waste may still be suitable for bioremediation using enzymatic activities, albeit without the economic incentives of a full recycling process. Here we develop a delivery system for PET hydrolases, based on engineered yeast biofilms that grow and secrete enzymes directly on the substrate to be degraded. The system does not require any enzyme purification steps, and the biofilms at least partially assimilate the degradation products into biomass. Our work provides proof-of-principle that demonstrates the potential of biofilm-based approaches.

bioengineering↗

bioRxiv2025

Near-cognate tRNAs dominate codon decoding times in simulated ribosomes

The codon sequence of messenger RNAs affects ribosome dynamics, translational control, and transcript stability. Here we describe an advanced computational modelling tool and its application to studying the effect of different tRNA species on the codon decoding process. We show that simulated codon decoding times are sensitive to the abundance of near-cognate tRNA species as well as cognate species, an aspect of the decoding system that is not fully considered in other computational modelling studies. We demonstrate that codon decoding times predicted by models that accurately define near-cognate tRNAs and that are parameterised with high-quality tRNA abundance datasets are highly similar to ribosome dwell times determined using experimental ribosome footprinting data, thereby confirming both the importance of near-cognate tRNAs for the codon decoding process and the general accuracy of our modelling tools.

systems biology↗

bioRxiv2021

Rpl24Bst mutation suppresses colorectal cancer by promoting eEF2 phosphorylation via eEF2K

Increased protein synthesis supports the rapid proliferation associated with cancer. The Rpl24Bst mutant mouse reduces the expression of the ribosomal protein RPL24 and has been used to suppress translation and limit tumorigenesis in multiple mouse models of cancer. Here we show that Rpl24Bst also suppresses tumorigenesis and proliferation in a model of colorectal cancer with two common patient mutations, Apc and Kras. In contrast to previous reports, Rpl24Bst mutation has no effect on ribosomal subunit abundance but suppresses translation elongation through phosphorylation of eEF2, reducing protein synthesis by 40% in tumour cells. Ablating eEF2 phosphorylation in Rpl24Bst mutant mice by inactivating its kinase, eEF2K, completely restores the rates of elongation and protein synthesis. Furthermore, eEF2K activity is required for the Rpl24Bst mutant to suppress tumorigenesis. This work demonstrates that elevation of eEF2 phosphorylation is an effective means to suppress colorectal tumorigenesis with two driver mutations. This positions translation elongation as a therapeutic target in colorectal cancer, as well as other cancers where the Rpl24Bst mutation has a tumour suppressive effect in mouse models.

cancer biology↗

bioRxiv2019

Hundreds of yeast mRNAs form a separately addressable, translation elongation-controlled regulon

In addition to the widespread and well documented control of protein synthesis by translation initiation, recent evidence suggests that translation elongation can also control protein synthesis rates. One of the proposed mechanisms leading to elongation control is the interference of slow ribosome movement around the start codon with efficient translation initiation. Here we estimate the frequency with which this mode of control occurs in bakers yeast growing in rich medium. Genome-wide data reveal that transcripts from around 20% of yeast genes show evidence of queueing ribosomes, which may be indicative of translation elongation control. Moreover, this subset of transcripts is sensitive to distinct regulatory signals compared to initiation-controlled mRNAs, and such distinct regulation occurs for example during the response to osmotic stress. Notethe previous version 2 of this preprint contained a Decision-Tree based analysis where we attempted to relate mRNA features to the presence or absence of queueing ribosome peaks. Since releasing that version, we performed additional controls for this analysis which strongly sugest that its results are random and should be ignored. Specifically, both the overall predictability and the importance of individual features is very similar for a real dataset where genes are labelled as containing a second SSU peak or not, and for a simulated control dataset containing an equal proportion of randomly labelled genes. We have removed this figure from the current version. The analysis together with the additional control remains accessible on the accompanying Github repository (github.com/tobiasvonderhaar/ribosomespeedcontrol) in the file "Figure 3 Obsolete (Classification).ipynb".

molecular biology↗

bioRxiv2018

Hidden patterns of codon usage bias across kingdoms

The genetic code is necessarily degenerate with 64 possible nucleotide triplets being translated into 20 amino acids. 18 out of the 20 amino acids are encoded by multiple synonymous codons. While synonymous codons are clearly equivalent in terms of the information they carry, it is now well established that they are used in a biased fashion. There is currently no consensus as to the origin of this bias. Drawing on ideas from stochastic thermodynamics we derive from first principles a mathematical model describing the statistics of codon usage bias. We show that the model accurately describes the distribution of codon usage bias of genomes in the fungal and bacterial kingdoms. Based on it, we derive a new computational measure of codon usage bias -- the distance [Formula] capturing two aspects of codon usage bias: (i) Differences in the genome-wide frequency of codons and (ii) apparent non-random distributions of codons across mRNAs. By means of large scale computational analysis of over 900 species across 2 kingdoms of life, we demonstrate that our measure provides novel biological insights. Specifically, we show that while codon usage bias is clearly based on heritable traits and closely related species show similar degrees of bias, there is considerable variation in the magnitude of [Formula] within taxonomic classes suggesting that the contribution of sequence-level selection to codon bias varies substantially within relatively confined taxonomic groups. Interestingly, commonly used model organisms are near the median for values of [Formula] for their taxonomic class, suggesting that they may not be good representative models for species with more extreme [Formula], which comprise organisms of medical an agricultural interest. We also demonstrate that amino acid specific patterns of codon usage are themselves quite variable between branches of the tree of life, and that some of this variability correlates with organismal tRNA content.

bioinformatics↗

bioRxiv2018

Experimental determination of codon usage-dependent selective pressure on high copy-number genes in Saccharomyces cerevisiae

One of the central hypotheses in the theory of codon usage evolution is that in highly expressed genes particular codon usage patterns arise because they facilitate efficient gene expression and are thus selected for in evolution. Here we use plasmid copy number assays and growth rate measurements to explore details of the relationship between codon usage, gene expression level, and selective pressure in Saccharomyces cerevisiae. We find that when high expression levels are required optimal codon usage is beneficial and provides a fitness advantage, consistent with evolutionary theory. However, when high expression levels are not required, optimal codon usage is surprisingly and strongly selected against. We show that this selection acts at the level of protein synthesis, and we exclude a number of molecular mechanisms as the source for this negative selective pressure including nutrient and ribosome limitations and proteotoxicity effects. These findings inform our understanding of the evolution of codon usage bias, as well as the design of recombinant protein expression systems.

molecular biology↗

bioRxiv2017

Codon-Dependent Translational Accuracy Controls Protein Quality in Escherichia coli but not in Saccharomyces cerevisiae

In order to generate a functional proteome, gene expression pathways must assemble proteins accurately according to the rules of the genetic code. General gene expression accuracy is known to be high, but errors nevertheless occur with measurable frequencies. Here we develop a mass-spectrometry (MS) based assay for the detection of a particular type of gene expression error, amino acid misincorporation. This assay allows assessing a much broader range of misincorporation events compared to current, very sensitive but also very specific enzyme reporter assays. Our assay uncovers a remarkably rich pool of error products for a model protein expressed in E. coli, which depend quantitatively on codon usage in the expression construct. This codon usage dependence can be explained in part as a function of the composition of the tRNA pool in this organism. We further show that codon-dependent differences in error levels correlate with measurable changes in specific protein activity. In contrast to E. coli, error levels are lower, and appear not to be codon usage dependent, when the same model protein is expressed in S. cerevisiae.

molecular biology↗