Genome browser & variant inspection
Load tracks and judge variants and coverage visually.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- IGV · UCSC Genome Browser
- Class GUI-need label
HY· Hybrid
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R08.06.01.02-T01 Classify five variants as real or artifact by read-level inspection in IGV
Original instruction
In IGV desktop load input/hg38 genome, input/sample.bam and input/calls.vcf. For each of the five variants in calls.vcf, inspect read support (strand balance, mapping quality, soft clips, position in read, duplicates) and label it 'true' or 'artifact' with a one-phrase reason in output/labels.csv.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Labels must match gold for all five (gold from orthogonal validation); reasons keyword-checked against the seeded defect per site (strand bias, low MAPQ, soft-clip cluster, duplicate stack).
Proposed partial credit
80% correct labels (16 pts each); 20% reasons matching the seeded causes.
R08.06.01.02-T02 Confirm a 12 kb heterozygous deletion in IGV and report its breakpoints
Original instruction
In IGV load input/sample.bam over chr7:117,100,000-117,400,000 with coverage and pair orientation colouring. Find the heterozygous deletion indicated by larger-than-expected insert sizes and halved coverage, and report its breakpoints to the nearest 100 bp plus supporting read-pair count in output/answer.json. Save the session as output/del.xml.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Both breakpoints within 200 bp of gold; supporting pair count within 20% of the reference count; session must show the BAM with insert-size colouring in the region. No SV VCF is provided. samtools, pysam and SV callers are not installed on the VM, so evidence must be read in IGV, and the session must show the inspected region.
Proposed partial credit
35% start breakpoint; 35% end breakpoint; 20% support count; 10% session saved.
R08.06.01.02-T03 Count canonical BRCA2 exons in the UCSC Genome Browser
Original instruction
In the UCSC Genome Browser (hg38) navigate to BRCA2 with GENCODE in full mode, identify the MANE Select transcript and report its exon count, chromosome span and strand in output/answer.json.
- Platforms
web· Any OS (web)- Applications
- UCSC Genome Browser
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollhover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Exon count, span (within 10 bp) and strand must equal the MANE Select record for BRCA2 in hg38.
Proposed partial credit
50% exon count; 30% span; 20% strand.
R08.06.01.02-T04 Export a TP53 locus browser view with SV and histone tracks as SVG
Original instruction
In the local UCSC mirror (Genome Browser in a Box, hg19) go to chr17:7,571,651-7,590,910 and add the tabix VCF and the bigWig served at the fixture URLs in input/urls.txt as custom tracks, set GENCODE to pack, hide default regulation tracks, and export the view as output/output.svg with both custom tracks visible.
- Platforms
web· Any OS (web)- Applications
- UCSC Genome Browser · Genome Browser in a Box
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP03· Visual layout- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
SVG must contain both custom track labels, the coordinate range in the title, a signal path whose peak x-positions match the bigWig maxima within 2% of the image width, and GENCODE in pack mode.
Proposed partial credit
30% both custom tracks present; 25% coordinates; 25% signal consistent with data; 20% track modes.
R08.06.01.02-T05 Inspect exome coverage and list exons needing resequencing in IGV
Original instruction
In IGV load input/exome.bam and input/panel_exons.bed over the 12 exons of the gene panel. Using coverage inspection, list in output/lowcov.csv every exon where coverage falls below 20× across more than 10% of its length, with the approximate low region.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Exon set must equal gold (computed with samtools depth); each reported low region must overlap the gold low region with IoU ≥0.5.
Proposed partial credit
60% exon set exact (scaled); 40% low regions localized.
R08.06.01.02-T06 Determine the phase of two nearby variants from read pairs in IGV
Original instruction
In IGV load input/trio_child.bam and view the two heterozygous variants listed in input/vars.txt (320 bp apart). Using reads and pairs that span both, decide whether they are in cis or trans, and report the call with the counts of reads supporting each haplotype in output/answer.json.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & click- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Phase call must equal gold (from long-read phasing); haplotype read counts within 15% of reference counts. No phased VCF or HP tags are present; phasing tools, samtools and pysam are absent from the image, and IGV's batch port is disabled.
Proposed partial credit
60% phase call; 40% supporting counts.
R08.06.01.02-T07 Save a UCSC session with a custom BED, GENCODE full and conservation hidden
Original instruction
In the UCSC Genome Browser (hg38) upload input/peaks.bed as a custom track, set it to pack, GENCODE to full, hide the Conservation track, navigate to chr8:127,700,000-127,760,000 and save a named session 'fc_session_01' in the sandbox account. Write the session URL to output/session.txt.
- Platforms
web· Any OS (web)- Applications
- UCSC Genome Browser
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Fetching the session settings must show the custom track in pack with the uploaded peaks, GENCODE full, Conservation hidden and the stated position (within 1 kb).
Proposed partial credit
30% custom track uploaded; 30% track visibility settings; 20% position; 20% session saved and shared.
R08.06.01.02-T08 Find the misannotated transcript isoform from RNA-seq coverage in IGV
Original instruction
In IGV load input/rnaseq.bam (with splice junctions), input/gencode.gtf and the four candidate transcripts of the gene in input/candidates.txt. Decide which annotated isoform the data actually support, and report it plus the two junctions that exclude the others in output/answer.json. Save output/isoform.xml.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Isoform ID must equal gold; both junctions must match gold coordinates within 2 bp and have ≥10 supporting reads in the BAM. Quantification and junction tools (salmon, StringTie, regtools), samtools and pysam are not installed in the image.
Proposed partial credit
50% isoform; 40% two junctions; 10% session saved.
R08.06.01.02-T09 Build a JBrowse 2 session with three tracks and export its session JSON
Original instruction
In JBrowse 2 (web, local server) with the hg38 assembly, add input/sample.cram, input/variants.vcf.gz and input/coverage.bw as tracks, open a linear genome view at chr12:25,200,000-25,260,000 with the alignments sorted by base at the KRAS G12 codon, and export the session to output/session.json.
- Platforms
web· Any OS (web)- Applications
- JBrowse 2
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
tutorial· Follow docs/tutorial- Proposed difficulty
- 3 / 5
Proposed verifier
session.json must declare the three tracks with the given adapters, the view locus within 1 kb, and a sort-by-base setting at the G12 position; re-loading the session in headless JBrowse must render all three tracks.
Proposed partial credit
40% three tracks; 25% locus; 25% sort setting; 10% session loads.
R08.06.01.02-T10 Identify the sample with a 3 Mb copy-number loss from segmented coverage tracks
Original instruction
In IGV load the four segmented copy-number files in input/cnv/ (sample1–4) with the hg38 genome. Find which sample carries a focal 3 Mb loss on chromosome 10 and report sample ID, approximate boundaries and log2 ratio in output/answer.json.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Sample ID exact; boundaries within 300 kb of gold; log2 ratio within 0.15. Distractor samples carry broad arm-level changes on the same chromosome.
Proposed partial credit
40% sample; 40% boundaries; 20% log2 ratio.
R08.06.01.02-T11 Compare ChIP-seq signal at a promoter across conditions in the WashU browser
Original instruction
In the WashU Epigenome Browser (hg38) load the three bigWigs in input/chip/ as a group with a shared y-axis fixed 0–30, add input/genes.bed, navigate to the MYOD1 promoter, and export a session link to output/session.txt plus a screenshot output/view.png. Report which condition has the highest promoter signal in output/answer.json.
- Platforms
web· Any OS (web)- Applications
- WashU Epigenome Browser
- Interfaces
brow· Browser- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scrolldrag· Drag & drop- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Session must hold the three tracks with fixed 0–30 scale and the locus; highest-signal condition must equal the gold computed over the promoter interval; screenshot shows three aligned tracks with equal scale (axis labels detected).
Proposed partial credit
30% tracks with fixed scale; 25% locus; 30% correct condition; 15% screenshot.
R08.06.01.02-T12 Confirm a de novo variant in a trio and report parental allele support
Original instruction
In IGV load input/child.bam, father.bam and mother.bam in a multi-locus panel at the candidate variant in input/candidate.txt. Decide whether it is de novo, inherited or an artifact, and report the call with child alt/ref counts and each parent's alt read count in output/answer.json.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & clickhover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Call must equal gold; counts within 10% of samtools mpileup values at that position with the same filters. One parent carries two low-quality alt reads that must not be called inherited. samtools, bcftools and pysam are not installed in the agent image, so allele counts must be read in IGV.
Proposed partial credit
45% call; 35% child counts; 20% parental counts.
R08.06.01.02-T13 Characterize a complex insertion from soft-clipped reads in IGV
Original instruction
In IGV load input/sample.bam at the locus in input/region.txt, show soft-clipped bases, and determine the inserted sequence (≥30 bp) and its exact insertion point from clipped read ends on both sides. Report the insertion point and sequence in output/answer.json.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & clickhover· Hover & context- Work product
WP12· Answer/decision- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Insertion point within 5 bp of gold; reported sequence ≥90% identical to gold over ≥30 bp (alignment). Assembly tools, samtools and pysam are not installed and no insertion call is in any VCF, so clipped sequence must be read in IGV.
Proposed partial credit
35% insertion point; 50% sequence identity; 15% consistent with reads on both sides.
R08.06.01.02-T14 Build a UCSC Track Hub for three bigWigs and load it as a composite
Original instruction
Create a UCSC track hub in output/hub/ (hub.txt, genomes.txt, trackDb.txt) serving the three bigWigs in input/bw/ for hg38 as a composite track with subgroups by mark, viewLimits 0:25 and a colour per mark. Serve output/hub/ at http://fixtures.local/hub/ (the harness web server) and load hub.txt in the local UCSC mirror (Genome Browser in a Box), verify it renders at chr1:1,000,000-1,100,000, and save the session link to output/session.txt.
- Platforms
web· Any OS (web)lin· Linux- Applications
- UCSC Genome Browser · Genome Browser in a Box
- Interfaces
brow· Browserterm· Terminal- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
type· Text & formsclick· Point & clickcode· Code & commands- Work product
WP08· App/system state- Conditions
tutorial· Follow docs/tutorialmultiapp· Multi-app- Proposed difficulty
- 4 / 5
Proposed verifier
hubCheck passes with no errors; trackDb declares a composite with three subtracks, subgroups, viewLimits 0:25 and distinct colours; the saved session shows the hub loaded and the three subtracks visible at the locus.
Proposed partial credit
35% valid hub files; 25% composite/subgroups; 20% display settings; 20% loaded session.
R08.06.01.02-T15 Judge four variants in homopolymers and low-complexity regions in IGV
Original instruction
In IGV load input/sample.bam and the four indels in input/indels.vcf, all inside homopolymers or short tandem repeats. Using read-level evidence (repeat length distribution, strand pattern, flanking mismatches) label each 'true' or 'artifact' in output/labels.csv with the repeat unit and length you observed.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Labels must match gold (from PCR-free long-read truth) for all four; repeat unit and length must match the reference annotation for each site. No repeat annotation track is provided; samtools, pysam and indel realigners are not installed in the image.
Proposed partial credit
80% correct labels (20 pts each); 20% repeat annotations correct.
R08.06.01.02-T16 Control: classify the five variants from pileup metrics on the command line
Original instruction
Control pair of T01. Using samtools and bcftools from the terminal on input/sample.bam and input/calls.vcf, compute per-site metrics (strand counts, MAPQ, read position, soft clips) and write 'true' or 'artifact' plus the deciding metric for each of the five variants to output/labels.csv.
- Platforms
mac· macOSlin· Linux- Applications
- samtools · bcftools
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Same labels as the T01 gold; the deciding metric must match the seeded defect per site. Scored as the CLI control of T01.
Proposed partial credit
80% correct labels (16 pts each); 20% deciding metrics.
R08.06.01.02-T17 Compare a locus between hg19 and hg38 in IGV and report the coordinate shift
Original instruction
In IGV load input/sample_hg19.bam with the hg19 genome and input/sample_hg38.bam with hg38, navigate both to the gene in input/gene.txt, and report the start coordinate of its first exon in each assembly plus the offset in output/answer.json. Note in output/notes.txt whether the region is inverted between assemblies.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickmulti· Cross-app- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Both coordinates within 10 bp of the gold annotations and the offset consistent; the inversion statement must match the UCSC liftOver chain for that region.
Proposed partial credit
35% hg19 coordinate; 35% hg38 coordinate; 30% orientation statement.
R08.06.01.02-T18 Review ten candidate somatic variants in a tumour/normal pair and triage them
Original instruction
In IGV load input/tumor.bam and input/normal.bam with input/candidates.vcf (10 sites). For each, label 'somatic', 'germline' or 'artifact' from read evidence in both samples, record tumour VAF, and write output/triage.csv. Save a snapshot per site to output/snapshots/ via IGV's snapshot command.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & clickhover· Hover & context- Work product
WP12· Answer/decision- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
≥9 of 10 labels must match gold; each tumour VAF within 0.05 of the pileup value; 10 snapshots present at the right loci (image centre coordinate recorded in filenames). samtools, pysam and VAF callers are not installed on the VM; snapshots must come from IGV's snapshot command.
Proposed partial credit
60% labels (6 pts per site); 25% VAFs; 15% snapshots.
R08.06.01.02-T19 Reconcile discordant variant calls from two pipelines in IGV
Original instruction
input/pipelineA.vcf and input/pipelineB.vcf disagree at eight sites in input/sample.bam. In IGV review each site and write output/adjudication.csv with the site, which pipeline is right (or 'neither'), and the evidence class (alignment error, left-alignment difference, filter threshold, true miss).
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & contextclick· Point & click- Work product
WP12· Answer/decision- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
≥7 of 8 adjudications must match gold and the evidence class must match the seeded cause at each site (two are representation-only differences that must be called 'neither'). bcftools, samtools and pysam are not installed in the image, so representation differences must be judged in the viewer.
Proposed partial credit
55% adjudications; 35% evidence classes; 10% complete file.
R08.06.01.02-T20 Check an exon for off-target CRISPR editing in amplicon reads with IGV
Original instruction
In IGV load input/amplicon.bam at the three predicted off-target sites in input/offtargets.bed. Decide for each whether indels at the cut site exceed the background seen in input/control.bam, and report edited yes/no with approximate indel frequency in output/answer.json.
- Platforms
mac· macOSwin· Windowslin· Linux- Applications
- IGV
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Calls must match gold at all three sites; frequencies within 3 percentage points of the pileup values. One site has a 1% sequencing-error background that must be called unedited. CRISPResso, samtools and pysam are not installed in the image.
Proposed partial credit
75% site calls (25 pts each); 25% frequencies.
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.