For stem cells to differentiate into the appropriate cell-type, transcription factors (TFs) must be tailored to highly specific expression patterns of lineage-specific genes. Abnormalities in the differentiation process can lead to consequences, such as impaired tissue regeneration and the development of certain cancers.
In a new study published in Cell titled, “Design principles of cell-state-specific enhancers in hematopoiesis,” researchers from the Centre for Genomic Regulation (CGR) at Barcelona Institute of Science and Technology have applied generative AI to design synthetic enhancers, regulatory DNA sequences that support transcription of a gene, that can effectively control gene expression in healthy mammalian cells by uncovering new gene regulation principles.
Led by Lars Velten, PhD, group leader at the CRG, the study designed enhancers with specificity to user-defined combinations of hematopoietic progenitor cell states from scratch as proof of concept.
“The potential applications are vast. It’s like writing software but for biology, giving us new ways of giving instructions to a cell and guiding how they develop and behave with unprecedented accuracy,” said Robert Frömel, graduate student in the Velten lab and first author of the study.
In blood stem cell differentiation, single-cell transcriptomics and imaging-based studies have demonstrated that key TFs are expressed in smooth, overlapping gradients throughout the stem and progenitor populations. However, the mechanism behind how enhancers transform these imbalances in an identical set of TFs into highly specific activity patterns remains an ongoing research question.
Genome-wide datasets of TF binding and chromatin accessibility have historically lacked functional information on the role of each TF in a given sequence context. Additionally, the natural genome is statistically underpowered to learn the rules by which it is decoded, as it only contains a few instances of the exponentially large number TF combinations and binding site arrangements.
To address this gap, the study leveraged 64,000 fully synthetic DNA sequences to dissect the design principles of cell-state specific enhancers during differentiation of blood stem cells to seven lineages.
“To create a language model for biology, you have to understand the language cells speak. We set out to decipher these grammar rules for enhancers so that we can create entirely new words and sentences,” said Velten.
For 38 key TF binding sites, the researchers characterized the function of each site in addition to the combinatorial effects of their interactions across blood stem cell differentiation. They found that enhancers containing binding sites of individual lineage factors, which drive differentiation in a specific direction, activated transcription. However, certain combinations of these sites were either inactive or acted as repressors, thereby creating enhancers that were sensitive to ratios in TF expression.
In this way, most enhancers acted as a volume dial, turning gene activity up or down, while certain combinations were on/off switches. The authors termed this concept as “negative synergy,” meaning two factors that usually turn a gene on individually could effectively shut that gene down when they occur together.
The results could provide a new tool for gene-therapy developers to boost or dampen the activity of genes in selectively targeted cells or tissues to make treatments more effective or reduce side effects.
The post AI Designed DNA Sequences Modulate Cell Differentiation for Gene Therapy Applications appeared first on GEN - Genetic Engineering and Biotechnology News.
In a new study published in Cell titled, “Design principles of cell-state-specific enhancers in hematopoiesis,” researchers from the Centre for Genomic Regulation (CGR) at Barcelona Institute of Science and Technology have applied generative AI to design synthetic enhancers, regulatory DNA sequences that support transcription of a gene, that can effectively control gene expression in healthy mammalian cells by uncovering new gene regulation principles.
Led by Lars Velten, PhD, group leader at the CRG, the study designed enhancers with specificity to user-defined combinations of hematopoietic progenitor cell states from scratch as proof of concept.
“The potential applications are vast. It’s like writing software but for biology, giving us new ways of giving instructions to a cell and guiding how they develop and behave with unprecedented accuracy,” said Robert Frömel, graduate student in the Velten lab and first author of the study.
In blood stem cell differentiation, single-cell transcriptomics and imaging-based studies have demonstrated that key TFs are expressed in smooth, overlapping gradients throughout the stem and progenitor populations. However, the mechanism behind how enhancers transform these imbalances in an identical set of TFs into highly specific activity patterns remains an ongoing research question.
Genome-wide datasets of TF binding and chromatin accessibility have historically lacked functional information on the role of each TF in a given sequence context. Additionally, the natural genome is statistically underpowered to learn the rules by which it is decoded, as it only contains a few instances of the exponentially large number TF combinations and binding site arrangements.
To address this gap, the study leveraged 64,000 fully synthetic DNA sequences to dissect the design principles of cell-state specific enhancers during differentiation of blood stem cells to seven lineages.
“To create a language model for biology, you have to understand the language cells speak. We set out to decipher these grammar rules for enhancers so that we can create entirely new words and sentences,” said Velten.
For 38 key TF binding sites, the researchers characterized the function of each site in addition to the combinatorial effects of their interactions across blood stem cell differentiation. They found that enhancers containing binding sites of individual lineage factors, which drive differentiation in a specific direction, activated transcription. However, certain combinations of these sites were either inactive or acted as repressors, thereby creating enhancers that were sensitive to ratios in TF expression.
In this way, most enhancers acted as a volume dial, turning gene activity up or down, while certain combinations were on/off switches. The authors termed this concept as “negative synergy,” meaning two factors that usually turn a gene on individually could effectively shut that gene down when they occur together.
The results could provide a new tool for gene-therapy developers to boost or dampen the activity of genes in selectively targeted cells or tissues to make treatments more effective or reduce side effects.
The post AI Designed DNA Sequences Modulate Cell Differentiation for Gene Therapy Applications appeared first on GEN - Genetic Engineering and Biotechnology News.