Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Detect m6A modifications from Oxford Nanopore direct RNA sequencing using m6Anet. Use when analyzing epitranscriptomic modifications from long-read RNA data without immunoprecipitation.
.claude/skills/bio-epitranscriptomics-m6anet-analysis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 9% | 0% |
<!--
#
#
-->
Documentation: https://m6anet.readthedocs.io/
bash# Basecall with Guppy (requires FAST5 files) guppy_basecaller \ -i fast5_dir \ -s basecalled \ --flowcell FLO-MIN106 \ --kit SQK-RNA002 # Align to transcriptome minimap2 -ax map-ont -uf transcriptome.fa reads.fastq > aligned.sam
pythonfrom m6anet.utils import preprocess from m6anet import run_inference # Preprocess: extract features from FAST5 preprocess.run( fast5_dir='fast5_pass', out_dir='m6anet_data', reference='transcriptome.fa', n_processes=8 ) # Run m6A inference run_inference.run( input_dir='m6anet_data', out_dir='m6anet_results', n_processes=4 )
bash# Preprocess m6anet dataprep \ --input_dir fast5_pass \ --output_dir m6anet_data \ --reference transcriptome.fa \ --n_processes 8 # Inference m6anet inference \ --input_dir m6anet_data \ --output_dir m6anet_results \ --n_processes 4
pythonimport pandas as pd results = pd.read_csv('m6anet_results/data.site_proba.csv') # Filter high-confidence m6A sites # probability > 0.9: High confidence threshold m6a_sites = results[results['probability_modified'] > 0.9]
<!-- AUTHOR_SIGNATURE: 9a7f3c2e-MD-BABU-MIA-2026-MSSM-SECURE -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→pass | 9,713 | 1,725 | -82% | 1 | 1 | 0% | 1,791 | 932 | -48% | 0 | 0 | — |
case-11 | fail→pass | 9,916 | 2,351 | -76% | 1 | 1 | 0% | 1,832 | 947 | -48% | 0 | 0 | — |
case-12 | pass→pass | 3,401 | 1,817 | -47% | 1 | 1 | 0% | 640 | 902 | +41% | 0 | 0 | — |
case-13 | pass→pass | 5,060 | 1,748 | -65% | 1 | 1 | 0% | 932 | 873 | -6% | 0 | 0 | — |
case-14 | pass→pass | 9,107 | 2,374 | -74% | 1 | 1 | 0% | 1,644 | 1,016 | -38% | 0 | 0 | — |
case-08 | pass→pass | 4,258 | 2,639 | -38% | 1 | 1 | 0% | 652 | 982 | +51% | 0 | 0 | — |
case-09 | pass→pass | 5,252 | 1,889 | -64% | 1 | 1 | 0% | 924 | 929 | +1% | 0 | 0 | — |
case-02 | fail→pass | 15,992 | 12,079 | -24% | 1 | 1 | 0% | 3,377 | 3,389 | +0% | 0 | 0 | — |
case-03 | fail→pass | 10,509 | 8,278 | -21% | 1 | 1 | 0% | 2,034 | 2,105 | +3% | 0 | 0 | — |
case-19 | pass→pass | 4,272 | 1,565 | -63% | 1 | 1 | 0% | 692 | 832 | +20% | 0 | 0 | — |
case-20 | fail→fail | 17,497 | 13,771 | -21% | 1 | 1 | 0% | 3,064 | 3,098 | +1% | 0 | 0 | — |
case-21 | fail→pass | 14,739 | 13,060 | -11% | 1 | 1 | 0% | 2,764 | 3,007 | +9% | 0 | 0 | — |
case-22 | fail→fail | 16,448 | 16,050 | -2% | 1 | 1 | 0% | 3,020 | 3,911 | +30% | 0 | 0 | — |
case-01 | fail→pass | 14,142 | 10,338 | -27% | 1 | 1 | 0% | 2,863 | 2,723 | -5% | 0 | 0 | — |
case-04 | pass→pass | 8,739 | 4,337 | -50% | 1 | 1 | 0% | 1,749 | 1,417 | -19% | 0 | 0 | — |
case-05 | fail→pass | 22,072 | 4,679 | -79% | 1 | 1 | 0% | 1,983 | 1,377 | -31% | 0 | 0 | — |
case-06 | fail→pass | 8,625 | 3,102 | -64% | 1 | 1 | 0% | 1,666 | 1,223 | -27% | 0 | 0 | — |
case-07 | fail→pass | 6,582 | 2,950 | -55% | 1 | 1 | 0% | 1,243 | 1,184 | -5% | 0 | 0 | — |
case-15 | pass→pass | 16,304 | 8,471 | -48% | 1 | 1 | 0% | 2,810 | 1,984 | -29% | 0 | 0 | — |
case-16 | pass→pass | 3,935 | 1,754 | -55% | 1 | 1 | 0% | 689 | 868 | +26% | 0 | 0 | — |
case-17 | fail→pass | 12,230 | 1,574 | -87% | 1 | 1 | 0% | 2,171 | 815 | -62% | 0 | 0 | — |
case-18 | fail→fail | 11,696 | 1,648 | -86% | 1 | 1 | 0% | 2,412 | 889 | -63% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +45 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/24/2026 | +52% |
Other measured skills in the registry, with their headline benchmark lift.