Sequencing analysis
Resolve the glycolytic transcription response
Read the task instructions and deliverables
# Resolve the glycolytic transcription response WM239A melanoma cells were exposed to trametinib for one hour. Characterize the response of the enzymes that convert glucose to pyruvate, and determine which individual enzyme-gene changes the reads support. Reconcile the overall pathway signal with its RNA sources. Use the catalytic enzyme families and candidate enzymes in the supplied reaction records for the ten core glycolytic steps, including their alternative gene products. Do not include pathway regulators or side reactions. The final comparison is the highest dose versus DMSO, keeping the workbook’s two replicate groups separate. Measure gene-span PRO-seq signal relative to the complete reference gene pool and give libraries equal weight. Report conservative source attribution relative to the supplied gene models. These are descriptive transcription measurements, not measurements of glucose consumption. The packet contains ten untrimmed, uniform three-million-read subsets, original sample records, genome-wide reference files and biochemical source passages. Choose your own analytical tools. Submit only answer.json using the provided fields.
How was it graded?
Code compares submitted answer fields with reference values. Written reasoning is not graded.
Fraction of criteria passed; every criterion has weight one.
LLM judge: None — deterministic grading
Task files
Choose an evidence file.
Grading criteria
Identify predominantly unresolved enzyme signals
Earns credit
The unresolved gene set is ALDOA and ALDOC. Their high-dose sense reads share annotated same-direction sources. PGK1 is not included: despite extensive annotation overlap, most observed reads resolve outside it.
Does not earn credit
The set is missing or differs after alias normalization.
Accepted alternatives
Order does not affect the set comparison. Repeated members, including aliases of the same member, fail the check. Leading and trailing whitespace is ignored. Accepted aliases: ENSG00000149925.22 → ALDOA; ENSG00000149925 → ALDOA; ENSG00000109107.14 → ALDOC; ENSG00000109107 → ALDOC.
Weight: 1
Input evidence
All 3,000,000 records. · Exact untrimmed experimental reads supporting source assignment and treatment comparison.
Reference evidence
exact-input-endpoints · variants, traps, gene_details and independent_checks.
Exact-input reference and defensible processing alternatives; source attribution is graded directly.
Recorded result / GPT 6 astra
Failed. unresolved_genes: matched 2 of 2 reference members. Extra: ENO3, ENO4, GAPDHS, PGAM2, PGK1.
Recover separately supported increases
Earns credit
The supported increase set contains ENO2, PFKM and PGAM1. HK3 and PFKL may be included or omitted because reasonable mapping and full-pool scaling choices change their borderline direction.
Does not earn credit
A required member is absent, an unsupported member such as ALDOC or TPI1 is present, or the field is missing.
Accepted alternatives
Order does not affect the set comparison. Repeated members, including aliases of the same member, fail the check. Leading and trailing whitespace is ignored. Accepted aliases: ENSG00000171314.9 → PGAM1; ENSG00000171314 → PGAM1; ENSG00000111674.9 → ENO2; ENSG00000111674 → ENO2; ENSG00000152556.17 → PFKM; ENSG00000152556 → PFKM. Optional members, ignored in the comparison: PFKL, HK3.
Weight: 1
Input evidence
All 3,000,000 records. · Exact untrimmed experimental reads supporting source assignment and treatment comparison.
Reference evidence
exact-input-endpoints · variants, traps, gene_details and independent_checks.
Exact-input reference and defensible processing alternatives; source attribution is graded directly.
Recorded result / GPT 6 astra
Passed. supported_increases: matched 3 of 3 reference members. No missing or extra members.
Measure the source-resolved fraction
Earns credit
In the submitted JSON, resolved_fraction must be approximately 0.896346, with an absolute tolerance of ±0.025 (inclusive). The displayed reference is rounded for readability; the original contract retains the exact value.
Does not earn credit
No credit for a missing field, a value that is not a finite JSON number, or a result outside the allowed tolerance.
Accepted alternatives
Conservative best-score assignments, high-MAPQ filtering, and tested alternative preprocessing are covered.
Weight: 1
Input evidence
All 3,000,000 records. · Exact untrimmed experimental reads supporting source assignment and treatment comparison.
Reference evidence
exact-input-endpoints · variants, traps, gene_details and independent_checks.
Exact-input reference and defensible processing alternatives; source attribution is graded directly.
Recorded result / GPT 6 astra
Failed. Submitted resolved_fraction = 0.66083; reference ≈ 0.896346. Absolute difference ≈ 0.235516; allowed difference ±0.025. Displayed numbers are rounded; the verdict is the original recorded result.
Measure the pathway response in the full reference pool
Earns credit
In the submitted JSON, pathway_response must be approximately 0.826955, with an absolute tolerance of ±0.1 (inclusive). The displayed reference is rounded for readability; the original contract retains the exact value.
Does not earn credit
No credit for a missing field, a value that is not a finite JSON number, or a result outside the allowed tolerance.
Accepted alternatives
The interval covers primary and high-confidence alignment pools, quality trimming and pooled versus equal-library descriptive summaries.
Weight: 1
Input evidence
All 3,000,000 records. · Exact untrimmed experimental reads supporting source assignment and treatment comparison.
Reference evidence
exact-input-endpoints · variants, traps, gene_details and independent_checks.
Exact-input reference and defensible processing alternatives; source attribution is graded directly.
Recorded result / GPT 6 astra
Passed. Submitted pathway_response = 0.823743; reference ≈ 0.826955. Absolute difference ≈ 0.00321183; allowed difference ±0.1. Displayed numbers are rounded; the verdict is the original recorded result.