Predicting 3D chromatin interactions from DNA sequence using Deep Learning

Piecyk, Robert S.; Schlegel, Luca; Johannes, Frank

doi:10.1016/j.csbj.2022.06.047

Cited by 14 publications

(7 citation statements)

References 75 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Combination of AI and more excellent single-cell omics derivative techniques is supposed to resolve this problem and reveal relationship between chromatin spatial structure and transcriptional regulation in single cells, especially single-cell Hi-C (scHi-C) 92,93 . Besides, many machine learning models have been applied to detect spatial structures at different scales of genome 94,95 , with promising usage in predicting TAD hierarchy. Furthermore, researchers of bioinformatics and genomics could make efforts to combine genome spatial architectures with better-fitting mathematical-physical models for higher sensitivity in interaction identification.…”

Section: Discussionmentioning

confidence: 99%

A comprehensive benchmarking with interpretation and operational guidance for the hierarchy of topologically associating domains

Xu,

Huang

et al. 2024

Nat Commun

View full text Add to dashboard Cite

Topologically associating domains (TADs), megabase-scale features of chromatin spatial architecture, are organized in a domain-within-domain TAD hierarchy. Within TADs, the inner and smaller subTADs not only manifest cell-to-cell variability, but also precisely regulate transcription and differentiation. Although over 20 TAD callers are able to detect TAD, their usability in biomedicine is confined by a disagreement of outputs and a limit in understanding TAD hierarchy. We compare 13 computational tools across various conditions and develop a metric to evaluate the similarity of TAD hierarchy. Although outputs of TAD hierarchy at each level vary among callers, data resolutions, sequencing depths, and matrices normalization, they are more consistent when they have a higher similarity of larger TADs. We present comprehensive benchmarking of TAD hierarchy callers and operational guidance to researchers of life science researchers. Moreover, by simulating the mixing of different types of cells, we confirm that TAD hierarchy is generated not simply from stacking Hi-C heatmaps of heterogeneous cells. Finally, we propose an air conditioner model to decipher the role of TAD hierarchy in transcription.

show abstract

Section: Discussionmentioning

confidence: 99%

A comprehensive benchmarking with interpretation and operational guidance for the hierarchy of topologically associating domains

Xu,

Huang

et al. 2024

Nat Commun

View full text Add to dashboard Cite

show abstract

“…To enhance accuracy and ensure our model’s comparability with similar deep learning frameworks (17, 28), we refined FitHiChIP’s original 5 kb anchor definitions to 2.5 kb regions centered on each anchor’s midpoint, to improve computational efficiency. Nucleotide sequences for these 2.5 kb windows were extracted from the Zea mays B73 reference genome (version 4) (26) for deep learning analysis.…”

Section: Methodsmentioning

confidence: 99%

“…Machine learning (ML) methods trained on 3C data in animals have emerged as powerful tools to dissect the DNA sequence grammar underlying chromatin biology. Deep Learning (DL) models, in particular, have demonstrated remarkable success in predicting 3D chromatin contacts directly from DNA sequences (17), with prediction accuracies in the range of 0.7 to 0.9 (17). The power of these models is that they can be combined with in silico mutagenesis, where arbitrary DNA sequence mutations are induced and evaluated for their impact on looping probabilities.…”

mentioning

confidence: 99%

GenomicLinks: Deep learning predictions of 3D chromatin loops in the maize genome

Schlegel,

Bhardwaj,

Shahryary

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Gene regulation in eukaryotes is partly shaped by the 3D organization of chromatin within the cell nucleus. Distal interactions between cis-regulatory elements and their target genes are widespread and many causal loci underlying heritable agricultural traits have been mapped to distal non-coding elements. The biology underlying chromatin loop formation in plants is poorly understood. Dissecting the sequence features that mediate distal interactions is an important step toward identifying putative molecular mechanisms. Here, we trained GenomicLinks, a deep learning model, to identify DNA sequence features predictive of 3D chromatin interactions in maize. We found that the presence of binding motifs of specific Transcription Factor classes, especially bHLH, are predictive of chromatin interaction specificities. Using an in silico mutagenesis approach we show the removal of these motifs from loop anchors leads to reduced interaction probabilities. We were able to validate these predictions with single-cell co-accessibility data from different maize genotypes that harbor natural substitutions in these TF binding motifs. GenomicLinks is currently implemented as an open-source web tool, which should facilitate its wider use in the plant research community.

show abstract

“…Wang et al [3] produced deep learning enables sophisticated analysis of complex plant genetic data. The deep learning approaches are more predictive than conventional methods Piecyk et al [4] they may be able to increase the accuracy of 3D chromatin interaction predictions. Deep learning algorithms are highly effective at capturing intricate non-linear correlations within data.…”

Section: Related Studiesmentioning

confidence: 99%

DNA based phenotype optimization of oryza sativa using machine learning and MolCNN

Soren,

Rajendran

2024

IJEECS

View full text Add to dashboard Cite

The prediction of phenotype from the genotype of oryza sativa (rice) is very crucial for optimizing the crop management. utilizing molecular convolutional neural networks (MolCNNs) and machine learning for crop management in oryza sativa provides a data-driven method for phenotype prediction based on DNA data, improving farming techniques. Data gathering, preparation, and integration of phenotypic and DNA data are all part of this process. Meaningful DNA features are extracted by MolCNN, and phenotypic qualities are predicted by a machine learning algorithm. Making educated decisions is ensured by assessing the model’s effectiveness, applying it to crop management, and updating it frequently. Choosing crop varieties, planting schedules, and management techniques are guided by molecular insights, which support sustainable agriculture and increase yields and quality. In the proposed research we are calculating pearson correlation coefficients between anticipated and actual trait values and the model’s performance on a test set. Additionally, it determines the (PCC) for every characteristic in the model and we have received a binary accuracy of 0.9998 in 139 seconds.

show abstract

Predicting 3D chromatin interactions from DNA sequence using Deep Learning

Cited by 14 publications

References 75 publications

A comprehensive benchmarking with interpretation and operational guidance for the hierarchy of topologically associating domains

A comprehensive benchmarking with interpretation and operational guidance for the hierarchy of topologically associating domains

GenomicLinks: Deep learning predictions of 3D chromatin loops in the maize genome

DNA based phenotype optimization of oryza sativa using machine learning and MolCNN

Contact Info

Product

Resources

About