Journal: Proceedings of the National Academy of Sciences of the United States of America

Article Title: Optimizing model representation for integrative structure determination of macromolecular assemblies

doi: 10.1073/pnas.1814649116

Figure Lengend Snippet: Comparison of representations of a large assembly. Comparing the performance of an approximately optimal representation (r′) with other uniform-resolution representations of 5 (r5), 30 (r30), and 50 (r50) residues per bead for the 10-protein transcription initiation and DNA repair factor TFIIH. Also shown is the performance of a five-residue per bead representation with the sampling time equal to that used for computing r′ and sampling the corresponding models (r5 [limited sampling]). Localization probability density maps specifying the probability of any volume element being occupied by a given bead in superposed good-scoring models (Top), EM map (Bottom, gray mesh), and representative models (Bottom, beads colored by protein) from the most populated cluster are shown for various representations. The total CPU time in seconds (t) for sampling models in various representations is shown as black bars (six-core dual Intel Xeon E5-2620 v3 processor). The total number of beads (n) for regions of unknown structure in each representation is shown in gray bars.

Article Snippet: The parameters used here are provided in SI Appendix , Table S1 A . Integrative modeling of the transcription/DNA repair factor TFIIH relied on a cryo-EM map of the complex, cross-links, X-ray structures, and comparative models of the constituent proteins ( 35 ).

Techniques: Sampling

Journal: Scientific Reports

Article Title: Thymocyte regulatory variant alters transcription factor binding and protects from type 1 diabetes in infants

doi: 10.1038/s41598-022-18296-4

Figure Lengend Snippet: rs138300818 G insertion allele—protecting from early-onset T1D—creates a thymocyte motif ( a ). At the same time, this motif disrupts the RFX5/7 binding motif which is present with the rs138300818 reference (null) allele ( b ). rs138300818 is located intronic in PTPRK . Non-coding RNA transcription was detected overlapping rs138300818 (grey vertical line) in all four studied thymocyte cell types (dark blue peaks). Roadmap Epigenomics chromatin state data for fetal thymus (PrimaryHMM) indicated strong transcription (state 4, green bar) overlapping the SNP flanking region ( c ). PCHiC data in fetal thymocytes and in naïve CD8 cells suggests that rs138300818 interacts with both PTPRK and THEMIS transcription start sites. The genetic association signal for early-onset T1D is colocalized with eQTL signal for THEMIS in human whole blood; rs138300818 reference allele—which forms the RFX7/5 binding motif—is associated with higher THEMIS expression ( d ).

Article Snippet: Roadmap Epigenomics fetal thymus ChromHMM model —based on histone modification data—predicted strong transcription overlapping rs138300818 (Fig. ).

Techniques: Binding Assay, Expressing

Journal: bioRxiv

Article Title: Denoising image-based spatial transcriptomics data with DenoIST

doi: 10.1101/2025.11.13.688387

Figure Lengend Snippet: a) Shown here is a region of the healthy lung sample assayed using Xenium, with the boundary expansion segmentation from 10x Xenium Ranger (Top) and Proseg (Bottom) segmentation. Each dot is a transcript molecule, molecules of lineage marker genes are coloured by their respective lineages. Black lines are segmentation boundaries. b) Log CPM (counts per million) normalised pseudobulked gene expression of 4 selected contaminating genes using 4 annotated immune cell types in the lung fibrosis Xenium data. Each data point is a sample and a lowess curve is fitted over all the samples. c) A high level schematic of the application of DenoIST.

Article Snippet: To address this research gap, we present DenoIST (Denoising Image-based Spatial Transcriptomics), a Poisson mixture model tailored for denoising IST data by reducing the effects of transcript contamination in downstream analysis tasks.

Techniques: Marker, Gene Expression

Journal: bioRxiv

Article Title: Denoising image-based spatial transcriptomics data with DenoIST

doi: 10.1101/2025.11.13.688387

Figure Lengend Snippet: Expression of ACTA2 from a zoomed-in section in Xenium human breast cancer dataset. Each dot is a segmented cell using 10x boundary expansion method, colour shows the log count of ACTA2 . Cells with 0 count are greyed out for visual clarity. b) Heatmap visualisation of gene expression of annotated cell types before (top) and after DenoIST (bottom). Columns are selected genes, annotated by the cell type they mark. Row are cell types. The log(mean count + 1) for each cell type is shown here. c) MECR before (top) and after (bottom) applying DenoIST to Xenium human breast cancer dataset. Rows and columns denote genes and each entry is the MECR of the corresponding pair. Note that genes that mark the same cell type are not expected to be mutually exclusive, but are shown here for positive control.

Article Snippet: To address this research gap, we present DenoIST (Denoising Image-based Spatial Transcriptomics), a Poisson mixture model tailored for denoising IST data by reducing the effects of transcript contamination in downstream analysis tasks.

Techniques: Expressing, Gene Expression, Positive Control

Journal: bioRxiv

Article Title: Denoising image-based spatial transcriptomics data with DenoIST

doi: 10.1101/2025.11.13.688387

Figure Lengend Snippet: a) UMAP visualisation of lung fibrosis data after applying DenoIST. Sample TILD028MA is shown here. Cells with 0 count are greyed out for visual clarity. b) An airway section from fibrotic sample VUILD110. Each dot is a cell. Cells with 0 count are greyed out for visual clarity. Annotated airway cell types and gene expression (raw counts and DenoIST-adjusted counts) for KRT5 and MUC5B are shown. c) Proportions of RCTD classification using raw counts and DenoIST-adjusted counts in healthy sample VUHD116A. d) RCTD assignment weights of the second highest lineage of each cell in healthy sample VUHD116A, stratified by their manually annotated lineages. Cells with a pure identity should have low weights for the incorrect lineages.

Article Snippet: To address this research gap, we present DenoIST (Denoising Image-based Spatial Transcriptomics), a Poisson mixture model tailored for denoising IST data by reducing the effects of transcript contamination in downstream analysis tasks.

Techniques: Gene Expression