EpiTrace.AI
Glossary

Epigenetics in plain language

56 terms that come up in methylation research, single-cell analysis, AI and research governance. Each entry gives the plain meaning first and the technical term second.

56 terms
#

5-hydroxymethylcytosine (5hmC)5-Hydroxymethylcytosin (5hmC)

An oxidised form of 5mC that appears when methylation is being removed; standard bisulfite methods cannot tell it apart from 5mC.

See also: 5-methylcytosine (5mC)Bisulfite conversion

A

AnnData (.h5ad)

A file format that stores a data matrix together with annotations for cells and features; EpiTrace.AI reads it.

See also: Cell × CpG matrix

ATAC-seq

A method that finds open, accessible regions of chromatin, where genes can be read.

See also: ChromatinMulti-omics

Attention

The part of a transformer model that weighs which inputs matter for a result; attention maps can help to explain it.

See also: TransformerExplainable AI (XAI)

B

β-valueβ-Wert

Share of methylated DNA copies at one CpG site, from 0 (none) to 1 (all).

See also: CpG siteM-value

Batch effectChargeneffekt

Differences caused by separate lab runs or devices rather than by biology.

See also: Quality control

Biomarker

A measurable feature, such as the methylation of a CpG site, that indicates a biological state; in research a candidate until validated.

See also: Marker scorecard

Bisulfite conversionBisulfit-Umwandlung

A chemical treatment that turns unmethylated C into U while methylated C stays; sequencing then shows methylation.

See also: EM-seqSingle-cell bisulfite sequencing

C

Cell × CpG matrixMatrix Zellen × CpG-Stellen

A table with one row per cell and one column per CpG site, holding methylation values.

See also: AnnData (.h5ad)

Chromatin

DNA packed with proteins in the cell nucleus; how tightly it is packed affects which genes can be read.

See also: ATAC-seq

Clustering

Grouping cells with similar profiles; common methods include Leiden and k-means.

See also: UMAP

CpG islandCpG-Insel

A DNA region with many CpG sites, often near the start of genes.

See also: CpG sitePromoter

CpG siteCpG-Stelle

A place in the DNA where the letter C is followed by G; the main location of methylation in mammals.

See also: β-valueCpG island

D

Data processing agreementAuftragsverarbeitung

A contract that governs how a service provider processes personal data on behalf of a research partner.

See also: PseudonymisationEthics vote

DifferentiationDifferenzierung

The process by which a less specialised cell becomes a specialised cell type.

See also: Stem cellProgenitor cell

DNA methylationDNA-Methylierung

Chemical tags on DNA that help to switch genes on or off without changing the DNA sequence.

See also: 5-methylcytosine (5mC)Epigenetics

E

EM-seq

Enzymatic methyl sequencing: enzymes replace bisulfite to reveal methylation, gentler on small DNA amounts.

See also: Bisulfite conversion

Epigenetic clockEpigenetische Uhr

A model that estimates age from the methylation of selected CpG sites.

See also: DNA methylation

EpigeneticsEpigenetik

Heritable changes in gene activity that do not alter the DNA sequence itself.

See also: DNA methylationChromatin

Epimutation (somatic)Epimutation (somatisch)

A random methylation change in a body cell that its daughter cells inherit; the basis of epigenetic lineage tracing.

See also: Lineage tracingClone

Ethics voteEthikvotum

Approval of a research project by an ethics committee before human data or samples are used.

See also: Data processing agreement

Evidence chainBelegkette

EpiTrace.AI's record of each step from sample to result: what was done, by whom, with which version and under which data rule.

See also: Reproducibility

Explainable AI (XAI)Erklärbare KI (XAI)

Methods that show why a model produced a result, so that researchers can check it.

See also: SHAP valuesAttention

F

FAIR principlesFAIR-Prinzipien

Guidelines for research data that is findable, accessible, interoperable and reusable.

See also: Reproducibility

Fate mappingSchicksalskartierung

Estimating which cell types the cells in a sample are heading towards.

See also: Cell fate

H

HaematopoiesisHämatopoese

The formation of blood cells from stem cells in the bone marrow; a key model system for lineage tracing.

See also: Stem cellClone

I

Imputation

Estimating missing values from the values that were measured.

See also: Sparsity

L

Lineage tracingLinienverfolgung

Reconstructing which cells descend from which, for example from inherited marks.

See also: Epimutation (somatic)Clone

M

M-valueM-Wert

A log-ratio transformation of the β-value that suits some statistical tests better.

See also: β-value

Marker scorecardMarker-Scorecard

EpiTrace.AI's ranking of CpG sites by how clearly they separate groups of cells.

See also: Biomarker

MethylomeMethylom

All methylation marks of a cell or sample.

See also: DNA methylation

Multi-omicsMultiomics

Measuring several molecular layers, such as methylation, gene activity and chromatin, in the same cells or samples.

See also: RNA-seqATAC-seqChIP-seq

P

Principal component analysis (PCA)Hauptkomponentenanalyse (PCA)

A linear method that finds the directions of largest variation in the data.

See also: Dimension reduction

Progenitor cellVorläuferzelle

A cell between a stem cell and a fully specialised cell.

See also: Stem cellDifferentiation

PromoterPromotor

The DNA region at the start of a gene where reading begins; its methylation often silences the gene.

See also: CpG islandEnhancer

PseudonymisationPseudonymisierung

Replacing names and direct identifiers with codes so that data can only be linked to a person with separately held information.

See also: Data processing agreement

Q

Quality controlQualitätskontrolle

Checks that remove low-quality cells and measurements before analysis.

See also: Batch effect

RNA-seq

Sequencing of RNA to measure which genes are active.

See also: Multi-omics

S

SHAP valuesSHAP-Werte

Numbers that show how much each input, such as a CpG site, contributed to a model result.

See also: Explainable AI (XAI)

Single-cell bisulfite sequencingEinzelzell-Bisulfit-Sequenzierung

Methods such as scBS-seq that read methylation across the genome of individual cells.

See also: Bisulfite conversionSparsity

SparsityDatenlücken

Single cells yield only some of their CpG sites, so many values are missing.

See also: Imputation

Spatial omicsRäumliche Omics

Measurements that keep the position of cells in the tissue.

See also: Multi-omics

T

t-SNE

A method that draws a 2D map of similar data points; used in the demo on this website.

See also: UMAPDimension reduction

Transcription factorTranskriptionsfaktor

A protein that binds DNA and switches genes on or off.

See also: EnhancerChIP-seq

Tumour heterogeneityTumorheterogenität

Differences between the cells of one tumour, for example in methylation, which single-cell methods make visible.

See also: Clone

V

Variational autoencoder (VAE)Variational Autoencoder (VAE)

A neural network that condenses each cell into a few meaningful numbers and can reconstruct it; methVAE applies this to methylation.

See also: Dimension reductionTransformer

No term matches your search.