Output
Overview
selscape organizes results by analysis type and population:
results/
├── positive_selection/
│ ├── selscan/
│ │ ├── {species}/1pop/{ppl}/{method}_{maf}/ # Within-population (iHS, nSL)
│ │ └── {species}/2pop/{pair}/{method}_{maf}/ # Cross-population (XP-EHH, XP-nSL)
│ └── scikit-allel/
│ ├── {species}/1pop/{ppl}/{method}/{window}_{step}/ # Negative Tajima's D
│ └── {species}/2pop/{pair}/{method}/{window}_{step}/ # Delta Tajima's D
├── balancing_selection/
│ ├── betascan/{species}/{ppl}/m_{core_frq}/
│ └── scikit-allel/{species}/{method}/{ppl}/{window}_{step}/ # Positive Tajima's D
├── dadi/{species}/dfe/{ppl}/
│ ├── InferDM/ # Demographic models
│ ├── InferDFE/ # DFE inference
│ ├── StatDFE/ # Confidence intervals
│ └── plots/ # Fitted model visualizations
Understanding File Paths
Throughout this documentation, file paths use the following wildcards:
{species}: Species name (e.g., "Human"){ppl}: Population code (e.g., "YRI"){pair}: Population pair (e.g., "YRI_CHS"){method}: Selection statistic (e.g., "ihs", "windowed_tajima_d"){maf}: Minor allele frequency (e.g., "0.05"){cutoff}: Top proportion threshold (e.g., "0.0005"){window}_{step}: Window and step sizes{ref_genome}: Reference genome build (e.g., "hg38"){demog}: Demographic model (e.g., "two_epoch"){dfe}: DFE model (e.g., "lognormal")
Positive Selection
selscan
Within-Population Statistics (iHS, nSL)
File path:
results/positive_selection/selscan/{species}/1pop/{ppl}/{method}_{maf}/{ppl}.normalized.{method}.maf_{maf}.top_{cutoff}.outliers.scores
Example output:
SNP CHR BP normalized_nsl abs_normalized_nsl
21:35702539 21 35702539 -4.88331 4.88331
20:57138931 20 57138931 4.68161 4.68161
21:28603177 21 28603177 4.63644 4.63644
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position |
| normalized_{method} | Normalized selection statistic (iHS or nSL) |
| abs_normalized_{method} | Absolute value of normalized statistic |
Positive values indicate unusually long haplotypes around the derived allele, whereas negative values indicate unusually long haplotypes around the ancestral allele. Large absolute values may be consistent with a recent selective sweep.
Cross-Population Statistics (XP-EHH, XP-nSL)
File path:
results/positive_selection/selscan/{species}/2pop/{pair}/{method}_{maf}/{pair}.normalized.{method}.maf_{maf}.top_{cutoff}.outliers.scores
Same format as within-population statistics.
Positive values indicate a stronger extended-haplotype signal in the first population (e.g., YRI in YRI_CHS), and negative values indicate a stronger signal in the second population (e.g., CHS in YRI_CHS).
More information: selscan.
scikit-allel
Windowed Tajima's D
File path:
results/positive_selection/scikit-allel/{species}/1pop/{ppl}/windowed_tajima/{window}_{step}/{ppl}.windowed_tajima..top_{cutoff}.outliers.scores
Example output:
SNP CHR BP tajima_d window_start window_end n_snps
20:61849 20 61849 -0.5400134 61849 161848 215
20:261849 20 261849 -0.2830888 261849 361848 400
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position (window center) |
| tajima_d | Tajima's D statistic |
| window_start | Window start position in base pairs |
| window_end | Window end position in base pairs |
| n_snps | Number of SNPs in window |
Negative values indicates an excess of rare variants, which can be consistent with a signature of positive selection, but may also be caused by demographic history (e.g., population expansion).
For all Tajima's D methods, each data point in the Manhattan plot represents the starting position (window_start) of a window.
Moving Tajima's D
File path:
results/positive_selection/scikit-allel/{species}/1pop/{ppl}/moving_tajima_d/{window}_{step}/{ppl}.moving_tajima_d.top_{cutoff}.outliers.scores
Similar format to windowed approach, but windows are defined by a fixed number of SNPs rather than base pairs.
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position (window center) |
| tajima_d | Tajima's D statistic |
| window_start | Window start position in base pairs |
| window_end | Window end position in base pairs |
| n_snps | Number of SNPs in window (constant) |
Delta Tajima's D (Cross-Population)
File path:
results/positive_selection/scikit-allel/{species}/2pop/{pair}/delta_moving_tajima_d/{window}_{step}/{pair}.delta_moving_tajima_d.top_{cutoff}.outliers.scores
Example output:
SNP CHR BP delta_tajima_d window_start window_end n_snps abs_delta_tajima_d
21:26659162 21 26659162 3.41144 26659162 26674022 100 3.41144
21:21456001 21 21456001 3.17360 21456001 21473170 100 3.17360
21:27242333 21 27242333 -3.07313 27242333 27258823 100 3.07313
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position (window center) |
| delta_tajima_d | Difference in Tajima's D between populations |
| window_start | Window start position in base pairs |
| window_end | Window end position in base pairs |
| n_snps | Number of SNPs in window |
| abs_delta_tajima_d | Absolute value of delta Tajima's D |
Large positive or negative delta Tajima’s D values indicate strong differences in the allele frequency spectrum between populations, which may be consistent with differential selection but can also be explained by demographic history. Here, windows are defined by a fixed number of SNPs.
More information: scikit-allel.
Balancing Selection
BetaScan
File path:
results/balancing_selection/betascan/{species}/{ppl}/m_{core_frq}/{ppl}.{ref_genome}.m_{core_frq}.b1.top_{cutoff}.outliers.scores
Example output:
SNP CHR BP B1
20:1914695 20 1914695 41.939494
20:1914712 20 1914712 41.479453
20:1914729 20 1914729 40.555049
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position |
| B1 | Beta1 statistic for balancing selection |
Larger B1 values indicate stronger evidence for long-term balancing selection (i.e., a balanced polymorphism maintained over extended periods).
More information: BetaScan.
scikit-allel
Windowed Tajima's D
File path:
results/balancing_selection/scikit-allel/{species}/windowed_tajima_d/{ppl}/{window}_{step}/{ppl}.windowed_tajima_d.top_{cutoff}.outliers.scores
Example output:
SNP CHR BP tajima_d window_start window_end n_snps
20:61849 20 61849 -0.5400134 61849 161848 215
20:261849 20 261849 -0.2830888 261849 361848 400
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position (window center) |
| tajima_d | Tajima's D statistic |
| window_start | Window start position in base pairs |
| window_end | Window end position in base pairs |
| n_snps | Number of SNPs in window |
Positive values indicates an excess of intermediate-frequency variants; it can be consistent with balancing selection, but can also reflect demographic history.
Moving Tajima's D
File path:
results/balancing_selection/scikit-allel/{species}/moving_tajima_d/{window}_{step}/{ppl}.moving_tajima_d.top_{cutoff}.outliers.scores
Similar format to windowed approach, but windows are defined by a fixed number of SNPs rather than base pairs.
| Column | Description |
|---|---|
| SNP | SNP identifier in format chromosome:position |
| CHR | Chromosome number |
| BP | Base pair position (window center) |
| tajima_d | Tajima's D statistic |
| window_start | Window start position in base pairs |
| window_end | Window end position in base pairs |
| n_snps | Number of SNPs in window (constant) |
Distribution of Fitness Effects
dadi/dadi-cli
Demographic Model Inference
File path: results/dadi/{species}/dfe/{ppl}/InferDM/{ppl}.{ref_genome}.{demog}.InferDM.bestfits
Example output:
# Log(likelihood) nu T misid theta
-29.738235902874877 3.2402258068957983 0.13580062396208237 0.0010470465761291449 321.237660524018
-29.73823629293304 3.2417998728753408 0.13578140459967458 0.0010576355371547388 321.23238379628344
-29.73823658527745 3.237659010918116 0.13591524984020853 0.0010605026099018222 321.22253191565875
| Column | Description |
|---|---|
| Log(likelihood) | Log-likelihood of the model fit (higher is better) |
| nu | Ratio of current population size to ancestral population size |
| T | Time since population size change, measured in 2Ne generations |
| misid | Proportion of misidentified ancestral states |
| theta | Population-scaled mutation rate (4Neμ) |
Demographic Model Fit
File path: results/dadi/{species}/dfe/{ppl}/plots/{ppl}.{ref_genome}.{demog}.fitted.png
The plot shows the observed frequency spectrum (data) compared to the model-predicted spectrum.
Distribution of Fitness Effects
File path: results/dadi/{species}/dfe/{ppl}/InferDFE/{ppl}.{ref_genome}.{demog}.{dfe}.InferDFE.bestfits
Example output:
# Log(likelihood) log_mu log_sigma misid theta
-28.737041385073574 2.5282728150148785 4.591179583286655 0.039371652920637594 742.0589958104816
-28.73804138508753 2.5282735377291865 4.591183662610013 0.0393715391668421 742.0589958104816
-28.737041388380618 2.5283038170363823 4.5912804442420105 0.039370931031321235 742.0589958104816
| Column | Description |
|---|---|
| Log(likelihood) | Log-likelihood of the DFE model fit (higher is better) |
| log_mu | Mean of log-normal distribution of selection coefficients |
| log_sigma | Standard deviation of log-normal distribution of selection coefficients |
| misid | Proportion of misidentified ancestral states |
| theta | Population-scaled mutation rate (4Neμ) |
DFE Model Fit Plot
File path: results/dadi/{species}/dfe/{ppl}/plots/{ppl}.{ref_genome}.{demog}.{dfe}.fitted.png
The plot shows the observed vs. model-predicted frequency spectrum for nonsynonymous mutations.
Mutation Proportion Plot
File path: results/dadi/{species}/dfe/{ppl}/plots/{ppl}.{ref_genome}.{demog}.{dfe}.fitted.mut.prop.png
The mutation proportion plot shows the proportions of deleterious mutations grouped by the population-scaled selection coefficient (2N|s|): nearly neutral mutations (< 1), weakly deleterious mutations (1–10), moderately deleterious mutations (10–100), and strongly deleterious mutations (> 100).
Confidence Intervals
File path: results/dadi/{species}/dfe/{ppl}/StatDFE/{ppl}.{ref_genome}.{demog}.{dfe}.godambe.ci
Example output:
Estimated 95% uncerts (theta adj), with step size 0.1): [2.45625971 3.17513531 0.11405905]
Lower bounds of 95% confidence interval : [-0.65471712 -0.74208004 -0.0046111 ]
Upper bounds of 95% confidence interval : [4.2578023 5.60819059 0.22350701]
Estimated 95% uncerts (theta adj), with step size 0.01): [2.19800876 2.55236006 0.10687941]
Lower bounds of 95% confidence interval : [-0.39646617 -0.11930478 0.00256854]
Upper bounds of 95% confidence interval : [3.99955135 4.98541534 0.21632737]
Estimated 95% uncerts (theta adj), with step size 0.001): [2.19575422 2.54701563 0.10682199]
Lower bounds of 95% confidence interval : [-0.39421163 -0.11396035 0.00262596]
Upper bounds of 95% confidence interval : [3.99729681 4.98007091 0.21626994]
Godambe-based 95% confidence intervals for all DFE parameters.
More information: dadi and dadi-cli.
Functional Annotation
Gene and functional annotation of outliers:
ANNOVAR Annotated Outliers
All outliers are annotated with gene information:
File path: {analysis_path}/{file_prefix}.annotated.outliers
Example output:
Chr Start End Ref Alt Func.refGene Gene.refGene ExonicFunc.refGene AAChange.refGene
20 14150039 14150039 C A intronic MACROD2 . .
20 17148100 17148100 C A intergenic OTOR;PCSK2 . .
20 23595497 23595497 G A intergenic CST9L;CST9 . .
| Column | Description |
|---|---|
| Chr | Chromosome number |
| Start | Variant start position in base pairs |
| End | Variant end position in base pairs |
| Ref | Reference allele |
| Alt | Alternative allele |
| Func.refGene | Functional region (exonic, intronic, intergenic, UTR, etc.) |
| Gene.refGene | Gene symbol(s) |
| ExonicFunc.refGene | Exonic function (synonymous, nonsynonymous, stopgain, stoploss, etc.) |
| AAChange.refGene | Amino acid change for coding variants |
Outlier Gene Lists
File path: {analysis_path}/{file_prefix}.outlier.genes
Example output:
Gene
HLCS
MACROD2
PCSK2
UQCC1
Simple list of unique genes extracted from annotated outliers.
More information: ANNOVAR.
GO Enrichment Analysis
Statistical testing for GO enrichment in outlier genes:
Gowinda Enrichment Results
File path: {analysis_path}/{file_prefix}.gowinda.enrichment.tsv
Example output:
GO_ID avg_genes_sim genes_found p_value p_adjusted genes_uniq genes_max genes_total description gene_list
GO:0034551 0.008 1 0.0082240000 1.0000000000 1 1 14 mitochondrial complex III uqcc1
GO:0031410 0.270 2 0.0247750000 1.0000000000 2 37 1017 cytoplasmic vesicle uqcc1,pcsk2
GO:0004077 0.030 1 0.0297130000 1.0000000000 1 1 1 biotin ligase activity hlcs
| Column | Description |
|---|---|
| GO_ID | Gene Ontology term identifier |
| avg_genes_sim | Average number of genes in simulations |
| genes_found | Number of outlier genes found in this GO term |
| p_value | Uncorrected p-value from simulation |
| p_adjusted | FDR-corrected p-value |
| genes_uniq | Number of unique genes in this GO term |
| genes_max | Maximum possible genes for this term |
| genes_total | Total genes in background set |
| description | GO term description |
| gene_list | Comma-separated list of outlier genes |
Enrichment Visualization
File path: {analysis_path}/{file_prefix}.gowinda.enrichment.png
The enrichment plot shows top 20 GO terms, sorted by their significance (FDR adjusted p-value) and then by the number of genes found. Bars are colored based on significance: red for highly significant enrichment (< 0.001), purple for significant enrichment (< 0.01), blue for moderately significant enrichment (< 0.05), and gray for non-significant results (≥ 0.05).
More information: Gowinda Tutorial.
Interactive HTML Reports
All methods generate interactive HTML tables with search, sort, and export functionality:
- Gene lists:
{analysis_path}/{file_prefix}.outlier.genes.html - Enrichment results:
{analysis_path}/{file_prefix}.gowinda.enrichment.html -
Best-fit parameters:
-
results/dadi/{species}/dfe/{ppl}/html/{ppl}.{ref_genome}.{demog}.InferDM.top10.bestfits.html results/dadi/{species}/dfe/{ppl}/html/{ppl}.{ref_genome}.{demog}.{dfe}.InferDFE.top10.bestfits.htmlresults/dadi/{species}/dfe/{ppl}/html/{ppl}.{ref_genome}.{demog}.{dfe}.godambe.ci.html