Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Identify differential m6A methylation between conditions from MeRIP-seq. Use when comparing epitranscriptomic changes between treatment groups or cell states.
.claude/skills/bio-epitranscriptomics-m6a-differential/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 20% | 0% |
<!--
#
#
-->
rlibrary(exomePeak2) # Define sample design # condition: factor for comparison design <- data.frame( condition = factor(c('ctrl', 'ctrl', 'treat', 'treat')) ) # Differential peak calling result <- exomePeak2( bam_ip = c('ctrl_IP1.bam', 'ctrl_IP2.bam', 'treat_IP1.bam', 'treat_IP2.bam'), bam_input = c('ctrl_Input1.bam', 'ctrl_Input2.bam', 'treat_Input1.bam', 'treat_Input2.bam'), gff = 'genes.gtf', genome = 'hg38', experiment_design = design ) # Get differential sites diff_sites <- results(result, contrast = c('condition', 'treat', 'ctrl'))
rlibrary(QNB) # Requires count matrices from peak regions # IP and input counts per sample qnb_result <- qnbtest( IP_count_matrix, Input_count_matrix, group = c(1, 1, 2, 2) # 1=ctrl, 2=treat ) # Filter significant # padj < 0.05, |log2FC| > 1 sig <- qnb_result[qnb_result$padj < 0.05 & abs(qnb_result$log2FC) > 1, ]
rlibrary(ggplot2) # Volcano plot ggplot(diff_sites, aes(x = log2FoldChange, y = -log10(padj))) + geom_point(aes(color = padj < 0.05 & abs(log2FoldChange) > 1)) + geom_hline(yintercept = -log10(0.05), linetype = 'dashed') + geom_vline(xintercept = c(-1, 1), linetype = 'dashed')
<!-- 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-01 | fail→fail | 16,399 | 13,396 | -18% | 1 | 1 | 0% | 3,379 | 3,378 | -0% | 0 | 0 | — |
case-13 | pass→pass | 6,893 | 3,577 | -48% | 1 | 1 | 0% | 1,501 | 1,364 | -9% | 0 | 0 | — |
case-14 | fail→pass | 6,476 | 2,536 | -61% | 1 | 1 | 0% | 1,272 | 1,088 | -14% | 0 | 0 | — |
case-15 | pass→pass | 10,949 | 2,453 | -78% | 1 | 1 | 0% | 2,070 | 1,117 | -46% | 0 | 0 | — |
case-07 | fail→pass | 12,219 | 3,999 | -67% | 1 | 1 | 0% | 2,279 | 1,356 | -41% | 0 | 0 | — |
case-02 | fail→fail | 15,483 | 11,813 | -24% | 1 | 1 | 0% | 3,500 | 3,311 | -5% | 0 | 0 | — |
case-03 | fail→fail | 18,089 | 9,950 | -45% | 1 | 1 | 0% | 3,674 | 2,797 | -24% | 0 | 0 | — |
case-04 | fail→pass | 17,630 | 13,877 | -21% | 1 | 1 | 0% | 3,319 | 3,404 | +3% | 0 | 0 | — |
case-05 | pass→pass | 11,762 | 8,455 | -28% | 1 | 1 | 0% | 2,336 | 2,232 | -4% | 0 | 0 | — |
case-06 | pass→pass | 14,746 | 12,303 | -17% | 1 | 1 | 0% | 2,738 | 2,993 | +9% | 0 | 0 | — |
case-08 | fail→pass | 11,395 | 6,806 | -40% | 1 | 1 | 0% | 2,085 | 1,739 | -17% | 0 | 0 | — |
case-09 | fail→fail | 8,774 | 8,845 | +1% | 1 | 1 | 0% | 1,608 | 2,150 | +34% | 0 | 0 | — |
case-10 | fail→pass | 8,591 | 5,717 | -33% | 1 | 1 | 0% | 1,507 | 1,811 | +20% | 0 | 0 | — |
case-11 | pass→pass | 9,360 | 6,534 | -30% | 1 | 1 | 0% | 1,991 | 2,083 | +5% | 0 | 0 | — |
case-12 | pass→pass | 7,743 | 5,594 | -28% | 1 | 1 | 0% | 1,613 | 1,813 | +12% | 0 | 0 | — |
case-16 | pass→pass | 13,984 | 3,608 | -74% | 1 | 1 | 0% | 2,012 | 1,341 | -33% | 0 | 0 | — |
case-17 | pass→fail | 7,428 | 2,387 | -68% | 1 | 1 | 0% | 1,172 | 1,070 | -9% | 0 | 0 | — |
case-18 | pass→pass | 8,933 | 2,586 | -71% | 1 | 1 | 0% | 1,538 | 1,144 | -26% | 0 | 0 | — |
case-19 | pass→fail | 4,469 | 1,761 | -61% | 1 | 1 | 0% | 785 | 915 | +17% | 0 | 0 | — |
case-20 | pass→pass | 10,146 | 3,950 | -61% | 1 | 1 | 0% | 1,894 | 1,442 | -24% | 0 | 0 | — |
case-21 | fail→pass | 10,420 | 3,739 | -64% | 1 | 1 | 0% | 1,952 | 1,381 | -29% | 0 | 0 | — |
case-22 | pass→pass | 5,943 | 2,602 | -56% | 1 | 1 | 0% | 1,080 | 1,077 | -0% | 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 +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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 | 7/26/2026 | +45% |
Other measured skills in the registry, with their headline benchmark lift.