HSeeker Contributors

7 files · 7.5 MB · gzip

Synthetic FASTA files used by the HSeeker benchmark suite (benchmarks/benchmark.py). Generated deterministically with fixed NumPy/Python random seeds documented in SEED_MANIFEST.json. Four sequence profiles: uniform (25 % each ACGT), ga_biased (90 % purine), ct_biased (90 % pyrimidine), realistic (~41 % GC). Two size tiers: small (~30 MB) and medium (~300 MB). Download via: python benchmarks/benchmark.py --from-zenodo

Kantor, Rose · Shakya, Migun · Ruth, Nelson · et al.

2,095 rows · 907 KB · fasta, tsv

A virus genome database representing 21,015 near-complete virus genomes collected from untargeted ultra-deep RNA/DNA combined sequencing of wastewater. Sequence data was provided by the CASPER consortium and raw data may be found on NCBI SRA under bioprojects PRJNA1247874 and PRJNA1198001. Data underwent read trimming, rRNA and human read removal, de novo assembly, and selection of high-quality viral contigs. Contigs were clustered at 95% identity and 85% query coverage to dereplicate. Chimera-checking required at least two independent assemblies of the same viral genome or presence of the genome in another reference database. Annotation made use of RdRpCATCH, geNomad, checkV, BLASTN against NCBI core-nt, and RNAVirHost. The RdRp fasta files contain representative RdRp sequences identified through homology to major RdRp reference databases and clustered at 90% sequence identity over 75% sequence coverage. Included sequences contain all three conserved RdRp motifs (A, B, and C) arranged in either the canonical ABC configuration or the permuted CAB configuration.

de Paula Oliveira, Hugo · Jacob Machado, Denis · Prieto Oliveira, Paula · et al.

1.4 MB

Appendix A - Supplementary Digital Materials These materials contain the supplementary digital materials for the manuscript " Evidence for recombination in dengue virus genomes ," written by Hugo de Paula Oliveira (ORCiD: 0000-0001-5842-7187), Denis Jacob Machado (ORCiD: 0000-0001-9858-4515), Paula Prieto-Oliveira (ORCiD: 0000-0001-6999-3221), and Kary Ocaña (ORCiD: 0000-0002-2151-7418) . All files use open, machine-readable formats (CSV, Markdown, plain text, FASTA, NEXUS, Python, Bash). Directory and file names contain no spaces or special characters. SupplementaryData1_Tables - Supplementary Tables See SupplementaryData1.tar.gz . Nine comma-separated value (CSV) files, one per supplementary table. SupplementaryTables.md describes the contents of each file. File Contents SupplementaryTables.md Brief description of each table and its columns. Table_S1.csv List of 6,906 DENV genomes retrieved from BV-BRC, including accession numbers and associated metadata (87 columns). Table_S2.csv Summary of software tools, versions, and parameter settings used in the analyses. Table_S3.csv Proportion of recombinants across clades and branch lengths from the phylogenetic tree. Table_S4.csv Taxa assigned to recombination events. Table_S5.csv Evolutionary models and parameter settings used to simulate alignments for RECOSIM performance evaluation. Table_S6.csv Distribution of recombinant DENV genomes per continent and country. Table_S7.csv Recombination events in DENV genomes detected both in this study and in previous published research. Table_S8.csv Evaluation of RECOSIM clustering strategies and parameter combinations. Table_S9.csv Comparison of RECOSIM and RDP5 detection performance. SupplementaryData2_Figures - Supplementary Figures See SupplementaryData2.tar.gz . A Markdown document with a contextual preamble and figure captions, together with eight PNG image files. Images are embedded in the Markdown and also available as standalone files. File Contents SupplementaryFigures.md Preamble describing the phylogenetic workflow and seven recombination case studies, followed by all figures with their captions. Figure_S1.png Conceptual workflow of the phylogenetic inference pipeline. Figure_S2.png SNP and similarity analysis for recombinant JQ922559 (DENV-4-I intra-genotypic event). Figure_S3.png SNP and similarity analysis for recombinant KX452048 (DENV-4-I intra-genotypic event). Figure_S4.png SNP and similarity analysis for recombinant MG560143 (DENV-2 inter-genotypic event). Figure_S5.png SNP and similarity analysis for recombinant KC964095 (DENV-2 inter-genotypic event). Figure_S6.png SNP and similarity analysis for recombinant KU517845 (DENV-2 inter-genotypic event). Figure_S7.png SNP and similarity analysis for recombinant JF295012 (DENV-3 inter-serotypic event). Figure_S8.png SNP and similarity analysis for recombinant ON123656 (DENV-1 inter-genotypic event). SupplementaryData3_Sequences - Sequence Data See SupplementaryData3.tar.gz . The multiple sequence alignment, genome partition scheme, and accession list for the DENV phylogenomic dataset. File Contents README.md Detailed description of each file in this directory. accessions.txt Plain-text list of 6,638 unique DENV genome accession IDs, one per line. alignment.fasta Multiple sequence alignment of DENV genomes in FASTA format, aligned against the FLAVi flavivirus reference framework. partitions.nexus Genome partition scheme in NEXUS format defining the boundaries of DENV genomic regions in FLAVi coordinates. tree.nwk Newick tree file of 6,642 unique flavivirus sequences generated with the phylogenetic inference pipeline. denv_zikv_wnv_alignment.fasta Multiple sequence alignment of DENV, ZIKV, and WNV genomes in FASTA format, aligned against the FLAVi flavivirus reference framework. SupplementaryData4_Scripts - Phylogenomic Analysis Scripts See SupplementaryData4.tar.gz . Python and Bash scripts supporting the phylogenomic analyses, with a Markdown methods description and a README. File Contents README.md Description of each script, including usage instructions and dependencies. PhylogenomicAnalyses.md Methods for the branch-length, phylogenetic placement, and tree topology analyses; includes accession IDs of the 192 sequences in the pruned DENV tree. ReproducePhylogenomicAnalyses.md Steps required to reproduce the phylogenomic analyses. Script_S1.py Phylogenetic inference pipeline integrating RAxML, TNT, and IQ-TREE2. Script_S2.py Extraction of recombinant proportions and branch lengths from phylogenetic trees. Script_S3.py Branch-length analysis: Mann-Whitney U test and Spearman rank correlation. Script_S4.py Clade-distance computation: smallest clade containing each recombinant and its parents. Script_S5.sh Visualisation of clade-size comparisons as a three-panel box-and-strip plot. Script_S6.sh Bash script generating TNT batch run files for the leave-one-out topology analysis. SupplementaryData5_Performance - RECOSIM Performance Evaluation See SupplementaryData5.tar.gz . File Contents performance.md Description of RECOSIM, the simulation design, evaluation protocol, and performance metrics (Hits ratio, Fails ratio, sensitivity, and precision).

Pérez Gallego, Ruth · von Meijenfeldt, F. A. Bastiaan · Bale, Nicole J. · et al.

2.0 MB

Abstract Paleontological and phylogenomic observations have shed light on the evolution of cyanobacteria. Nevertheless, the emergence of heterocytes, specialized cells for nitrogen fixation, remains unclear. Heterocytes are surrounded by heterocyte glycolipids (HGs), which contribute to protection of the nitrogenase enzyme from oxygen. Here, by comprehensive HG identification and screening of HG biosynthesis genes throughout cyanobacteria, we identify HG analogs produced by specific and distantly related non-heterocytous cyanobacteria. These structurally less complex molecules probably acted as precursors of HGs, suggesting that HGs arose after a genomic reorganization and expansion of ancestral biosynthetic machinery, enabling the rise of cyanobacterial heterocytes in an increasingly oxygenated atmosphere. Subsequently, HG chemical structure evolved convergently in response to environmental pressures. Our results open a new chapter in the potential use of diagenetic products of HGs and HG analogs as fossils for reconstructing the evolution of multicellularity and division of labor in cyanobacteria. Here we supply: Supplementary Data 1. Selected cyanobacterial genomes from the PATRIC genome database (now part of the BV-BRC database). Files called 'selected_Cyanogenomes.genome_*.20220430.txt' are sourced from the PATRIC File Transfer Protocol server (ftp.patricbrc.org). 'gtdbtk.bac120.summary.tsv' is the GTDB-Tk output file, and 'qa.summary_extended.txt' the CheckM output file. Supplementary Data 2. HG biosynthetic gene clusters in selected PATRIC genomes and 14 newly sequenced genomes. The file 'islands_on_contigs.3_ORFs_in_between.expanded_island_with_nucleotide_positions.txt' contains the location of all hits to Anabaena sp. PCC 7120 HG biosynthesis genes. ORFs were predicted with Prodigal. The structure of a contig is as follows: "genome | contig". The structure of a hit is as follows: "ORF number on contig | query ( e -value; bit-score; start of alignment in query; end of alignment in query; query coverage per subject; start of alignment in subject; end of alignment in subject; subject coverage) [nucleotide position on contig start; nucleotide position on contig end; direction]". Non-overlapping hits on the same ORF (see Online Methods) are connected with '&&&' characters. An asterisk ('*') indicates that the hit is located at most three ORFs from a contig edge. Clusters of hits that are at most three open reading frames (ORFs) apart are connected with '~~~' characters. The file 'Supplementary_table.script_1.txt' contains a summary of all identified hgl islands (i.e. clusters containing at least seven unique HG biosynthesis gene hits). Supplementary Data 3. HG biosynthetic gene clusters in 255,388 prokaryotic genomes from the PATRIC genome database (now part of the BV-BRC database). The file called 'PATRIC_20230120.selection_c50_c10.txt' contains information on the selected PATRIC genomes based on data sourced from the PATRIC File Transfer Protocol server ( ftp.patricbrc.org ). The file 'all_tree_of_life_genomes.islands_on_contigs.3_ORFs_in_between.expanded_island_with_nucleotide_positions.txt' contains the location of all hits to Anabaena sp. PCC 7120 HG biosynthesis genes. ORFs were predicted with Prodigal. The structure of a contig is as follows: "genome | contig". The structure of a hit is as follows: "ORF number on contig | query ( e -value; bit-score; start of alignment in query; end of alignment in query; query coverage per subject; start of alignment in subject; end of alignment in subject; subject coverage) [nucleotide position on contig start; nucleotide position on contig end; direction]". Non-overlapping hits on the same ORF (see Online Methods) are connected with '&&&' characters. An asterisk ('*') indicates that the hit is located at most three ORFs from a contig edge. Clusters of hits that are at most three open reading frames (ORFs) apart are connected with '~~~' characters. Supplementary Data 4. Phylogeny of representative cyanobacterial genomes based on a core gene superalignment. The folder contains the files used to generate Fig. 2a. The directory 'IQ-TREE' contains the tree file and iTOL annotation files. The file 'dRep.representative_to_cluster.txt' contains the dRep clusters. Note that the manually defined subclades in the iTOL annotation file 'iTOL_annotation.manually_defined_clades.DATASET_STYLE.txt' have a different numbering from the paper: subclades 0 and 1 are the 'heterocytous sister clades', and subclades 2-10 in the annotation file are heterocytous subclades 1-9 in the paper, respectively. Supplementary Data 5. Lipid data files. The folder contains all the UHPLC-HRMS n (Orbitrap) datafiles used in this study. The directory 'CCY strains' includes 24 heterocytous cyanobacterial cultures corresponding to 23 strains grown in nitrogen-deficient media, the resulting data are shown in Supplementary Table 10. Directory 'HglT mutant' contains the datafiles used to generate Supplementary Table 15. The directory 'LEGE strains' includes the UHPLC-HRMS n (Orbitrap) and GC-MS datafiles corresponding to eight cultures of two non-heterocytous strains grown in media with and without nitrogen for 38 to 77 days, the resulting data are shown in Supplementary Tables 10, 17 and 18. Supplementary Data 6. Plasmid maps. GenBank and FASTA files of plasmids generated in this study. 'HglT deletion' directory contains the genomic region surrounding hglT in the wild-type strain and after deletion used to generate Supplementary Fig. 15. pAM5404 is shown in Supplementary Fig. 16 and p(A)RP0XX are shown in Supplementary Fig. 17. Supplementary Data 7. Phylogenies of seven hgl island genes and of a concatenated alignment of these genes. The folder contains the files used to generate Supplementary Fig. 18 (in the directory 'gene_trees_hgl_islands'), and Fig. 4 and related figures (in the directory 'gene_trees_hgl_islands_4'. The directories contain the alignments and trimmed alignments, IQ-TREE output files, and iTOL annotation files. The file 'gene_trees_hgl_islands/analysis_individual_gene_trees/explore_individual_gene_clusters.ipynb' contains the code to identify the five hgl islands that contain genes with incongruent evolutionary histories. Supplementary Data 8. Phylogeny of hglE A homologs. The folder contains the files used to generate Supplementary Fig. 24 and related figures. The file 'selected_hglE_hits.txt' contains the selected hglE A hits and the genomic cluster on which they are located. The folder contains the alignment and trimmed alignment, IQ-TREE output files, and iTOL annotation files. All the code used in this publication including scripts used for: genome assemblies, download of genomes from public repositories, quality and contamination checks, genome analysis, construction of the phylogenetic trees, hgl island identification, etc. The shell script 'commands.sh' within each directory contains all the code used to generate the content in the directory. All the figures used in this publication including the figures in the Supplementary Information file.

Finan, Chris

2.5 MB

Variant mapping reference files used by the variant-mapper package to assign dbSNP rsIDs and VEP-derived annotation metadata to variants in a text file. Common biallelic variants, VEP-annotated, provided for both GRCh37 (b37) and GRCh38 (b38) with tabix indexes.

Papadopoulou, Sofia · Langenheder, Silke · Lindström, Eva

778 KB

This dataset contains raw and processed data generated as part of in vitro transplant experiments conducted using water collected from the catchment area of lake Siggeforasjön, central Sweden. Amplicon sequencing (16S rRNA genes and transcripts): Operational taxonomic unit (OTU) table (otu_table.txt). Associated taxonomy table containing assignments for all OTUs (taxonomy_table.txt). OTU sequences and bootstrap support values generated using the DADA2 taxonomic assignment are provided for each taxonomic level. Raw reads are available through the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB108836. Flow cytometry: Flow cytometry measurements were used to quantify bacterial abundance and the relative proportions of high nucleic acid (HNA) to low nucleic acid (LNA) bacterial particles, expressed as the HNA/LNA ratio. Raw flow cytometry standard (.FCS) files from all analyzed samples are provided (FCS_files.zip). Bacterial protein production: Raw and processed bacterial protein production measurements based on leucine incorporation (Protein_production_raw_data.xlsx). The workbook includes: (i) raw measurements for samples, positive and negative controls; (ii) calculation steps used to derive bacterial protein production rates; and (iii) descriptions of all variables included in the calculations. Metadata table: Metadata file linking sample identifiers across datasets (Metadata.csv). Includes correspondence between sample name, sequencing read filenames (.fastq) deposited in ENA and associated flow cytometry (.FCS) files. Column descriptions are provided in Metadata_README.txt. Water flow: Water flow data (in m³ s⁻¹), originally obtained from the S-HYPE hydrological model through the Vattenwebb service of the Swedish Meteorological and Hydrological Institute. Daily data correspond to subcatchment SUBID 8110 (Siggeforasjön) for 2022. Files S1-S3: These supplementary files present relative abundance patterns of stream and lake communities across DNA- and RNA-based datasets (bacterial phylum or OTU level).

Fianco, Marcos · Melo, Gabriel A. R.

75 rows · 30 KB

Aligned Fasta files from Fianco & Melo (in press) Evolution of the false-leaf katydids (Orthoptera: Tettigoniidae: Phaneropterinae): molecular phylogeny and divergence times

Marques, João Pedro · Nevado, Bruno · Sim-Sim, Manuela

6.1 MB

This repository contains the repeat and gene annotation files associated with the chromosome-level genome assembly of Corema album (ENA accession GCA_964187925.1). Four GFF3 files are provided: Gene annotation (GFF3) Gene annotation with embedded FASTA sequence records Repeat annotation (GFF3) Repeat annotation with embedded FASTA sequence records These files represent the final structural annotations generated for the Corema album reference genome and are provided as a permanent public archive complementing the assembly available through the European Nucleotide Archive (ENA). Genome assembly accession: GCA_964187925.1 BioProject accession: PRJEB76683

Niz, José María · Sauka, Diego

309 KB

This dataset contains supporting bioinformatic outputs, genome annotation files, and supplementary sequence data derived from the genome assembly of Bacillus thuringiensis serovar entomocidus strain BGSC 4I1 (HD-10). The genome assembly is publicly available through the European Nucleotide Archive (ENA) under accession GCA_984526485 (WGS accession CFGYKP010000000.1; BioProject PRJEB113047; BioSample SAMEA122677898). The dataset provides reproducibility-oriented files generated during genome characterization, including genome annotation outputs and results from specialized bioinformatic analyses used for taxonomic assignment, insecticidal protein identification, and secondary metabolite biosynthetic gene cluster prediction. In addition, complete coding sequences of insecticidal protein genes identified in the genome are provided in FASTA format. The interpretation and contextualization of these analyses are described in the associated genome report deposited in Zenodo. This dataset serves as a companion resource to the associated genome report and facilitates the reproducibility and reuse of the genomic analyses presented therein.

Lopez-Neblina, Fernando

59 MB

This repository contains all datasets generated and analyzed in support of the manuscript: "Hierarchical territorial organization of IL1B intron 3 reveals compartmentalized intronic regulatory architectures." The repository includes: • IL1B intron 3 FASTA sequences • Selected territorial microdomains (Z4C2, Z2C4, and Z2C6) • TFBS analyses generated using FIMO and JASPAR • RNA-binding protein interaction analyses • RNAfold outputs • RNAstructure outputs • trRosettaRNA structural reconstructions • Three-dimensional RNA models (*.pdb) • Supplementary tables • Manuscript and associated figures These materials are provided to ensure full reproducibility of the analyses described in the associated manuscript.

Hubbard, Alfred · Solares, Edwin · Hemming-Schroeder, Elizabeth

88 rows · 18 KB · fasta

These are the files needed to run the Broad Institute's malaria amplicon pipeline for the PvGAP Plasmodium vivax panel, described in detail here . They consist of FASTA files containing the forward and reverse primers and another FASTA file containing reference sequences for each target, derived from the PvP01 reference genome.

Kim, Kwanho · Lin, Zechuan · Simmons, Sean · et al.

1 files · 37 KB · pdf

This dataset contains CCS corrected HiFi long-read DNA sequencing (lrDNAseq) in FASTQ format for 100 PMDBS samples from Parkinson's patients and healthy controls. It's part of the PD5D atlas, where the same subjects were also profiled with other omics assays including genotyping, single-cell ATACseq, and spatial transcriptomics.

Kim, Kwanho · Lin, Zechuan · Simmons, Sean · et al.

1 files · 37 KB · pdf

This dataset contains raw FASTQ files from the midbrain single-nucleus RNA sequencing (snRNAseq) dataset with hybrid selection for the matching PMDBS samples from the PD5D chort. The same subjects were also profiled with other omics assays including genomic DNAseq, genotyping, single-cell ATACseq, and spatial transcriptomics.

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 37 KB · pdf

This dataset consists of raw sequencing snRNA-seq data (10x Genomics Chromium Next GEM Single Cell 3ʹ). The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocol followed to isolate nuclei from postmortem brain samples and to prepare sequencing libraries can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a single sequencing library. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors.

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 36 KB · pdf

This dataset consists of raw sequencing snRNA-seq data using ParseBio Evercode Whole Transcriptome. The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocol followed to isolate nuclei from postmortem brain samples and to prepare sequencing libraries can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a specific ParseBio barcode. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors.

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 36 KB · pdf

This dataset consists of raw sequencing snATAC-seq data (10x Genomics Chromium Next GEM Single Cell ATAC v2). The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocol followed to isolate nuclei from postmortem brain samples and to prepare sequencing libraries can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a single sequencing library. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors. (edited)

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 36 KB · pdf

This dataset consists of raw sequencing snATAC-seq data (HyDrop v2). The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocol followed to isolate nuclei from postmortem brain samples can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a single sequencing library. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors.

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 36 KB · pdf

This dataset consists of raw sequencing ATAC-seq data (Scale-ATAC pre-indexing followed by 10x Genomics snATAC v2). The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocol followed to isolate nuclei from postmortem brain samples and to prepare sequencing libraries can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a single sequencing library. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors.

Pančíková, Alexandra · Theunis, Koen · Hulselmans, Gert · et al.

1 files · 36 KB · pdf

This dataset consists of raw sequencing ATAC-seq data (Scale-ATAC + HyDrop v2). The data is part of an overall set of samples derived from postmortem midbrain (n=140), cingulate cortex (n=190) and motor cortex (n=4) of healthy donors (n=114), patients with Parkinson's disease (n=75) or patients with other neurological disorder (n=1). The protocols followed to isolate nuclei from postmortem brain samples and to prepare sequencing libraries can be found below. To increase throughput and to decrease batch effects, several donors have been pooled together into a single sequencing library. To computationally demultiplex the nuclei to their corresponding donors, cellsnp-lite (version commit: aad18644adcde853c313362a856a24245c9b91f7) followed by vireo (https://github.com/single-cell-genetics/vireo/pull/108 ) has been used. The population VCF with the donor genotypes derived from whole genome sequencing data has been used to assign nuclei back to their donors.