Why Gene Maps Matter
A noncoding variant can overlap a promoter or enhancer without lying inside the gene it may regulate. To propose a target gene, the variant’s location therefore has to be connected to a gene map: a representation of gene bodies, transcripts, transcription start sites, and, where available, assigned regulatory elements. These maps help organize evidence, but a map-based candidate is not automatically the gene through which a variant affects a trait.
Three kinds of evidence are useful to keep distinct. A variant may directly overlap a promoter assigned to a gene; it may overlap another regulatory element, such as an enhancer, that is linked to a gene; or it may be near a gene by a distance rule. Each supports a different step in the argument. Direct overlap locates the variant in an annotated element, a link connects an element to a gene, and distance provides a positional clue. None alone proves a functional effect.
Reading a Gene Map
A gene can have several transcripts and therefore several TSSs. A promoter assignment may be made for each transcript, or a resource may select one representative TSS. The same genomic region can consequently overlap a promoter assigned to one transcript but not another. When interpreting an assignment, check which transcript and which promoter definition the map uses; a promoter window is an operational annotation, not a universal boundary shared by all genes.
The strand matters when describing which direction is upstream of a TSS: upstream is toward lower coordinates for a plus-strand transcript and toward higher coordinates for a minus-strand transcript. However, the ordinary absolute distance between a variant and a TSS does not depend on strand. For comparisons in this tutorial, coordinates must refer to the same genome assembly, and interval endpoints must use the same convention. We use zero-based, half-open intervals: an interval \([a,b)\) includes positions from \(a\) up to, but not including, \(b\). A single-base variant at coordinate \(x\) is represented as \([x,x+1)\).
Three Routes from a Variant to a Candidate Gene
For a promoter-overlap assignment, the first question is whether the variant’s base falls in the promoter interval specified by the map. If it does, the variant directly overlaps an element assigned to that gene or transcript. This is more specific than merely being close to a TSS, but it still does not establish that the variant changes promoter activity or affects the trait.
For an enhancer-to-gene assignment, the reasoning has two steps. First, determine whether the variant overlaps an annotated enhancer. Second, check whether that enhancer is linked to a gene in the relevant map or resource. A link might be supported by a particular experimental or computational procedure; its score and interpretation depend on that source. A link score should not be called a probability unless it has specifically been calibrated and defined as one. If an enhancer has links to multiple genes, preserve those alternatives rather than silently choosing one.
Distance-based evidence asks how close the variant is to a gene, often by distance to the TSS. It is useful for describing local genomic context and for generating candidates when stronger links are unavailable. It is not a substitute for an element-to-gene link: regulatory elements can act on genes that are not the nearest by a chosen distance rule. “Nearest gene” also depends on the rule—nearest TSS and nearest gene body need not identify the same gene.
Worked Example: A Variant Overlapping an Assigned Promoter
Suppose a hypothetical variant is at coordinate 500,100. In a gene map on the same assembly, Gene A has a plus-strand transcript with TSS 500,050 and an assigned promoter interval \([499,900,500,200)\). Gene B has a TSS at 500,700 and no promoter interval covering the variant. Does the map support a direct promoter-overlap candidate?
The variant occupies \([500,100,500,101)\). Since \(499,900 \leq 500,100 < 500,200\), this base lies inside Gene A’s assigned promoter interval. The TSS distance is
Thus Gene A is supported by direct overlap with its annotated promoter. Gene B is 600 bp from the variant’s coordinate to its TSS, since \(|500{,}100-500{,}700|=600\) bp, but that proximity is not promoter overlap. The result supports proposing Gene A as a promoter-linked candidate; it does not show that the variant alters transcription or that Gene A mediates a trait association.
Checking Element Overlap and Following a Link
Overlap can be checked directly from interval endpoints. For a variant base \([x,x+1)\) and an element interval \([a,b)\), overlap occurs exactly when the intervals share at least one coordinate. For this single-base variant, the condition is \(a \leq x < b\). For two general intervals \([a,b)\) and \([c,d)\), the length of their intersection is \(\max(0,\min(b,d)-\max(a,c))\). This expression is zero if they are disjoint and positive if they overlap.
Worked Example: An Enhancer Link Favors a Non-nearest Candidate
In a hypothetical locus, a variant lies at 2,000,050 and an annotated enhancer occupies \([2,000,000,2,000,120)\). The enhancer-to-gene resource lists a link score of 0.82 for Gene B and 0.31 for Gene A. Their TSS coordinates are 2,001,000 and 1,999,800, respectively. Check the overlap, compare distances, and interpret the link scores.
The variant base is \([2,000,050,2,000,051)\), and \(2,000,000 \leq 2,000,050 < 2,000,120\), so it overlaps the enhancer. The TSS distances are
Gene A is nearer by TSS distance, but the supplied enhancer-link score is higher for Gene B. If the resource defines larger scores as stronger links, the map therefore gives stronger enhancer-link support to Gene B than to Gene A, despite Gene B being farther away. The values 0.82 and 0.31 are scores as given; without a calibration definition, they are not probabilities that either gene is the true target. Both genes remain candidates for follow-up, and the score difference alone does not establish a causal target.
Comparing Evidence Without Collapsing It
These evidence types answer different questions, so a clear interpretation reports them separately. A direct promoter overlap identifies a variant inside an assigned promoter. An enhancer link proposes a gene for an overlapping enhancer. A distance calculation describes spatial proximity. When the clues agree, the assignment is more coherent, but agreement is not independent confirmation if the annotations share underlying data or assumptions. When they disagree, the disagreement should be reported rather than resolved automatically by choosing the nearest gene or the largest uncalibrated score.
| Evidence | What is checked | What it supports | What it does not establish |
|---|---|---|---|
| Direct promoter overlap | Variant base falls inside a promoter interval assigned to a gene or transcript | The variant is located in an annotated promoter for that gene or transcript | A regulatory effect, a trait mechanism, or a unique target gene |
| Enhancer-to-gene link | Variant overlaps an element, and that element has a recorded link to a gene | The linked gene is a proposed target of the overlapping element | That the link score is a probability or that the variant changes the element |
| Distance | Distance to a TSS or gene-body boundary under a stated rule | Relative genomic proximity | A functional connection or the identity of the affected gene |
Worked Example: Distance as a Clue, Not a Link
Consider a hypothetical noncoding variant at coordinate 3,000,500. Gene X has a TSS at 2,999,900, and Gene Y has a TSS at 3,002,000. The map contains no promoter overlap or enhancer-to-gene link for this variant. Which gene is nearest by TSS distance, and what conclusion is warranted?
For Gene X,
For Gene Y,
Gene X is the nearer gene by the stated TSS rule. Because no direct promoter overlap or element-to-gene link is provided, the map supports only a distance-based candidate designation for Gene X. It does not show that Gene X is the target; Gene Y or another gene could be relevant through regulatory evidence not present in this map.
A Practical Interpretation Sequence
Confirm the variant and gene-map annotations use the same genome assembly and coordinate convention. Represent the variant consistently as a base or interval before testing overlap.
Identify the transcript or TSS used for each promoter annotation. Record whether the variant base overlaps the stated promoter interval; do not infer overlap from closeness alone.
If the variant overlaps an enhancer, list the genes linked to that element and report the resource’s scores according to their stated scale, without treating an uncalibrated score as a probability.
State whether distance is measured to the TSS or to the gene body, give the result in base pairs, and use “nearest” only relative to that explicit rule.
Separate observed overlap, recorded links, and positional proximity. Name candidate genes and explain which evidence supports each, without presenting an annotation as proof of causation.
Common Mistake: Treating the Nearest Gene as the Target
A tempting but incorrect argument is: “The variant is closest to Gene A, so Gene A must be its target.” The calculation may correctly identify Gene A as nearest by TSS, but distance alone does not establish a regulatory connection. A variant can overlap an element linked to a more distant gene, and gene-map annotations can also be incomplete or dependent on transcript choice. The correction is to report the nearest gene as a distance-based candidate and keep that evidence distinct from promoter overlap and enhancer-to-gene links.
A second caution is to avoid describing a gene as “the target” when the available evidence only nominates it. A disciplined conclusion might say, “The variant overlaps a promoter assigned to Gene A,” or “The overlapping enhancer has a stronger recorded link score for Gene B than for Gene A.” Those statements accurately describe the map. A causal claim would require evidence beyond positional annotation.
Check Your Understanding
Use the definitions and distinctions in this tutorial to answer the following.
- A variant at coordinate 700,020 is compared with a promoter interval \([700,000,700,050)\). Does the variant base overlap that interval under the stated half-open convention?
- Why can a variant be nearer to one gene’s TSS while an enhancer-overlap annotation supports a different gene?
- What is the difference between distance to a TSS and distance to a gene body?
- A resource reports an enhancer-to-gene score of 0.9. What must be known before interpreting this value as a probability?
- Write one cautious sentence describing a variant that is nearest to Gene C but has no promoter overlap or enhancer-to-gene link in the available map.