Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Bio Alignment Validation · FreedomIntelligence | DecimalAI

Bio Alignment Validation

SkillScore100Maintained

bio-alignment-validation

Validate alignment quality with insert size distribution, proper pairing rates, GC bias, strand balance, and other post-alignment metrics. Use when verifying alignment data quality before variant calling or quantification.

scienceby FreedomIntelligence

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Data Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/bio-alignment-validation/SKILL.md

decimalai skills pull bio-alignment-validation --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/bio-alignment-validation/SKILL.md https://app.decimal.ai/s/bio-alignment-validation/SKILL.md

Raw file: /s/bio-alignment-validation/SKILL.md ↗ · machine-readable summary at /s/bio-alignment-validation.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 50d ago

Yes — +27% lift · at higher cost+132% tokens0% turns100% of whole cases pass with skill (22/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	136%	0%
case-07	✗→✓	▲ Improved	98%	0%
case-12	✗→✓	▲ Improved	128%	0%
case-15	✗→✓	▲ Improved	114%	0%
case-18	✗→✓	▲ Improved	59%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/1/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

FreedomIntelligence / OpenClaw-Medical-Skills

skills/bio-alignment-validation/SKILL.md

View on GitHub ↗

Imported 8/1/2026Version v2

Skill contents

SKILL.md

<!--

COPYRIGHT NOTICE

This file is part of the "Universal Biomedical Skills" project.

Copyright (c) 2026 MD BABU MIA, PhD <md.babu.mia@mssm.edu>

All Rights Reserved.

This code is proprietary and confidential.

Unauthorized copying of this file, via any medium is strictly prohibited.

Provenance: Authenticated by MD BABU MIA

-->

Alignment Validation

Post-alignment quality control to verify alignment quality and identify issues.

Insert Size Distribution

Insert size should match library preparation protocol.

samtools stats

bash
samtools stats input.bam > stats.txt
grep "^IS" stats.txt | cut -f2,3 > insert_sizes.txt

Picard CollectInsertSizeMetrics

bash
java -jar picard.jar CollectInsertSizeMetrics \
    I=input.bam \
    O=insert_metrics.txt \
    H=insert_histogram.pdf

Expected Insert Sizes

| Library Type | Expected Size | |--------------|---------------| | Standard WGS | 300-500 bp | | PCR-free | 350-550 bp | | RNA-seq | 150-300 bp | | ChIP-seq | 150-300 bp | | ATAC-seq | Multimodal |

Python Insert Size Analysis

python
import pysam
import numpy as np
import matplotlib.pyplot as plt

def get_insert_sizes(bam_file, max_reads=100000):
    sizes = []
    bam = pysam.AlignmentFile(bam_file, 'rb')
    for i, read in enumerate(bam.fetch()):
        if i >= max_reads:
            break
        if read.is_proper_pair and not read.is_secondary and read.template_length > 0:
            sizes.append(read.template_length)
    bam.close()
    return sizes

sizes = get_insert_sizes('sample.bam')
print(f'Median insert size: {np.median(sizes):.0f}')
print(f'Mean insert size: {np.mean(sizes):.0f}')
print(f'Std dev: {np.std(sizes):.0f}')

plt.hist(sizes, bins=100, range=(0, 1000))
plt.xlabel('Insert Size')
plt.ylabel('Count')
plt.savefig('insert_size_dist.pdf')

Proper Pairing Rate

Percentage of reads correctly paired.

samtools flagstat

bash
samtools flagstat input.bam

samtools flagstat input.bam | grep "properly paired"

Calculate Pairing Rate

bash
proper=$(samtools view -c -f 2 input.bam)
mapped=$(samtools view -c -F 4 input.bam)
rate=$(echo "scale=4; $proper / $mapped * 100" | bc)
echo "Proper pairing rate: ${rate}%"

Expected Rates

| Metric | Good | Marginal | Poor | |--------|------|----------|------| | Proper pair | > 90% | 80-90% | < 80% | | Mapped | > 95% | 90-95% | < 90% | | Singletons | < 5% | 5-10% | > 10% |

GC Bias

GC content correlation with coverage.

Picard CollectGcBiasMetrics

bash
java -jar picard.jar CollectGcBiasMetrics \
    I=input.bam \
    O=gc_bias_metrics.txt \
    CHART=gc_bias_chart.pdf \
    S=gc_summary.txt \
    R=reference.fa

deepTools computeGCBias

bash
computeGCBias \
    -b input.bam \
    --effectiveGenomeSize 2913022398 \
    -g hg38.2bit \
    -o gc_bias.txt \
    --biasPlot gc_bias.pdf

Interpret GC Bias

| Issue | Symptom | |-------|---------| | Under-representation | Low GC coverage drops | | Over-representation | High GC coverage elevated | | PCR bias | Strong correlation |

Strand Balance

Forward and reverse strand should be balanced.

Calculate Strand Ratio

bash
forward=$(samtools view -c -F 16 input.bam)
reverse=$(samtools view -c -f 16 input.bam)
echo "Forward: $forward"
echo "Reverse: $reverse"
ratio=$(echo "scale=4; $forward / $reverse" | bc)
echo "F/R ratio: $ratio"

Check Strand Bias per Chromosome

bash
for chr in chr1 chr2 chr3; do
    fwd=$(samtools view -c -F 16 input.bam $chr)
    rev=$(samtools view -c -f 16 input.bam $chr)
    echo "$chr: F=$fwd R=$rev ratio=$(echo "scale=2; $fwd/$rev" | bc)"
done

Mapping Quality Distribution

Extract MAPQ Distribution

bash
samtools view input.bam | cut -f5 | sort -n | uniq -c | sort -k2 -n

Calculate Mean MAPQ

bash
samtools view input.bam | awk '{sum+=$5; count++} END {print "Mean MAPQ:", sum/count}'

MAPQ Thresholds

| MAPQ | Meaning | |------|---------| | 0 | Multi-mapper | | 1-10 | Low confidence | | 20-30 | Moderate | | 40+ | High confidence | | 60 | Unique (BWA) |

Chromosome Coverage Balance

Calculate Per-Chromosome Coverage

bash
samtools idxstats input.bam | awk '{print $1, $3/$2}' | head -25

Check for Aneuploidy/Contamination

bash
samtools idxstats input.bam | awk '$3 > 0 {
    sum += $3
    len[$1] = $2
    reads[$1] = $3
} END {
    for (chr in reads) {
        expected = len[chr] / sum * reads[chr]
        ratio = reads[chr] / expected
        if (ratio < 0.8 || ratio > 1.2) print chr, ratio
    }
}'

Mismatch Rate

Picard CollectAlignmentSummaryMetrics

bash
java -jar picard.jar CollectAlignmentSummaryMetrics \
    I=input.bam \
    R=reference.fa \
    O=alignment_summary.txt

Key Metrics

| Metric | Description | Good Value | |--------|-------------|------------| | PCT_PF_READS_ALIGNED | Mapped % | > 95% | | PF_MISMATCH_RATE | Mismatches | < 1% | | PF_INDEL_RATE | Indels | < 0.1% | | STRAND_BALANCE | Strand ratio | ~0.5 |

Comprehensive Validation Script

bash
#!/bin/bash
BAM=$1
REF=$2
NAME=$(basename $BAM .bam)
OUTDIR=${3:-qc}

mkdir -p $OUTDIR

echo "=== Alignment Validation: $NAME ===" | tee $OUTDIR/report.txt

echo -e "\n--- Flagstat ---" | tee -a $OUTDIR/report.txt
samtools flagstat $BAM | tee -a $OUTDIR/report.txt

echo -e "\n--- Mapping Rate ---" | tee -a $OUTDIR/report.txt
mapped=$(samtools view -c -F 4 $BAM)
total=$(samtools view -c $BAM)
rate=$(echo "scale=2; $mapped / $total * 100" | bc)
echo "Mapping rate: ${rate}%" | tee -a $OUTDIR/report.txt

echo -e "\n--- Proper Pairing ---" | tee -a $OUTDIR/report.txt
proper=$(samtools view -c -f 2 $BAM)
pair_rate=$(echo "scale=2; $proper / $mapped * 100" | bc)
echo "Proper pairing: ${pair_rate}%" | tee -a $OUTDIR/report.txt

echo -e "\n--- Insert Size ---" | tee -a $OUTDIR/report.txt
samtools stats $BAM | grep "insert size average" | tee -a $OUTDIR/report.txt

echo -e "\n--- Strand Balance ---" | tee -a $OUTDIR/report.txt
fwd=$(samtools view -c -F 16 $BAM)
rev=$(samtools view -c -f 16 $BAM)
strand_ratio=$(echo "scale=3; $fwd / $rev" | bc)
echo "Forward: $fwd, Reverse: $rev, Ratio: $strand_ratio" | tee -a $OUTDIR/report.txt

echo -e "\n--- Chromosome Coverage ---" | tee -a $OUTDIR/report.txt
samtools idxstats $BAM | head -25 | tee -a $OUTDIR/report.txt

echo -e "\nReport: $OUTDIR/report.txt"

Python Validation Module

python
import pysam
import numpy as np
from collections import Counter

class AlignmentValidator:
    def __init__(self, bam_file):
        self.bam = pysam.AlignmentFile(bam_file, 'rb')

    def mapping_rate(self):
        stats = self.bam.get_index_statistics()
        mapped = sum(s.mapped for s in stats)
        unmapped = sum(s.unmapped for s in stats)
        return mapped / (mapped + unmapped) * 100

    def proper_pair_rate(self, sample_size=100000):
        proper = 0
        paired = 0
        for i, read in enumerate(self.bam.fetch()):
            if i >= sample_size:
                break
            if read.is_paired:
                paired += 1
                if read.is_proper_pair:
                    proper += 1
        return proper / paired * 100 if paired > 0 else 0

    def mapq_distribution(self, sample_size=100000):
        mapqs = []
        for i, read in enumerate(self.bam.fetch()):
            if i >= sample_size:
                break
            if not read.is_unmapped:
                mapqs.append(read.mapping_quality)
        return Counter(mapqs)

    def strand_balance(self, sample_size=100000):
        forward = 0
        reverse = 0
        for i, read in enumerate(self.bam.fetch()):
            if i >= sample_size:
                break
            if not read.is_unmapped:
                if read.is_reverse:
                    reverse += 1
                else:
                    forward += 1
        return forward / (forward + reverse) if (forward + reverse) > 0 else 0.5

    def report(self):
        print(f'Mapping rate: {self.mapping_rate():.1f}%')
        print(f'Proper pairing: {self.proper_pair_rate():.1f}%')
        print(f'Strand balance: {self.strand_balance():.3f}')

        mapq_dist = self.mapq_distribution()
        high_qual = sum(v for k, v in mapq_dist.items() if k >= 30)
        total = sum(mapq_dist.values())
        print(f'High MAPQ (>=30): {high_qual/total*100:.1f}%')

    def close(self):
        self.bam.close()

validator = AlignmentValidator('sample.bam')
validator.report()
validator.close()

Quality Thresholds Summary

| Metric | Good | Warning | Fail | |--------|------|---------|------| | Mapping rate | > 95% | 90-95% | < 90% | | Proper pairing | > 90% | 80-90% | < 80% | | Duplicate rate | < 10% | 10-20% | > 20% | | Strand balance | 0.48-0.52 | 0.45-0.55 | Outside | | Mean MAPQ | > 40 | 30-40 | < 30 | | GC bias | < 1.2x | 1.2-1.5x | > 1.5x |

Related Skills

bam-statistics - Basic flagstat and depth
duplicate-handling - Mark/remove duplicates
alignment-filtering - Filter low-quality reads
chip-seq/chipseq-qc - ChIP-specific metrics

Illustrative Want to see what bio-alignment-validation would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/2/2026, 16:04 UTC · 22 cases

Lift

+27%

100% vs 73%

Agent turns

1.0 vs 1.0

Tokens

+132%

4,197.05 vs 1,805.86

fail → pass · 6pass → pass · 16

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I have raw paired-end FASTQ files sample_R1.fastq.gz and sample_R2.fastq.gz straight off the Illumina sequencer. How do I generate quality score distribution boxplots and per-base sequence content summaries for raw FASTQ reads before performing genome alignment?	pass→pass	9,780	12,760	+30%	1	1	0%	1,936	5,702	+195%	0	0	—
case-02 I have a coordinate-sorted and duplicate-marked BAM file sample.bam ready for small variant discovery. What command line parameters do I pass to GATK HaplotypeCaller to perform germline SNP and indel calling against reference hg38.fasta?	pass→pass	9,397	5,895	-37%	1	1	0%	1,835	4,173	+127%	0	0	—
case-03 I just received an unsorted BAM file raw_alignments.bam from an aligner run. I need to sort the reads by genomic coordinates and generate an index file sorted_alignments.bam.bai. How do I accomplish this coordinate sorting and indexing?	pass→pass	7,002	4,775	-32%	1	1	0%	1,382	4,021	+191%	0	0	—
case-04 I am writing a bash command to count properly paired reads in sample.bam using samtools view. A colleague suggested using -F 2 to keep properly paired reads. What is the correct samtools view flag to filter for reads that are properly paired?	pass→pass	5,959	5,862	-2%	1	1	0%	1,149	4,070	+254%	0	0	—
case-05 I need to count mapped reads in alignments.bam using samtools view. Someone told me -f 4 counts mapped reads. Which flag correctly filters out unmapped reads to return only mapped reads?	pass→pass	6,787	4,934	-27%	1	1	0%	1,271	4,021	+216%	0	0	—
case-06 I am calculating strand balance in sample.bam using samtools view. I want to separate reverse strand reads from forward strand reads. My teammate claims flag 4 isolates reverse strand reads. Which flag should I use with samtools view to filter for reverse strand reads, and which flag excludes them for forward strand reads?	fail→pass	9,440	27,514	+191%	1	1	0%	1,759	4,149	+136%	0	0	—
case-07 I am conducting post-alignment validation on two libraries: a standard WGS library and an ATAC-seq library. I expected both to show a single sharp peak at 150 bp. What is the expected insert size range for standard WGS, and what shape should the ATAC-seq insert size distribution exhibit?	fail→pass	11,984	5,991	-50%	1	1	0%	2,087	4,126	+98%	0	0	—
case-08 When validating alignment insert sizes for RNA-seq and ChIP-seq sequencing libraries in BAM outputs, what is the standard expected insert size range? A baseline default of 500-800 bp was suggested.	pass→pass	14,034	6,273	-55%	1	1	0%	2,517	4,203	+67%	0	0	—
case-09 I need to run Picard to calculate insert size metrics on sample.bam and generate a visual plot PDF named insert_hist.pdf. What is the Picard tool name and the specific parameter flag for setting the output PDF histogram path?	pass→pass	2,472	2,589	+5%	1	1	0%	460	3,568	+676%	0	0	—
case-10 I want to measure GC bias across my BAM alignment input.bam using Picard against the human reference genome hg38.fa. Should I use Picard GcMetrics with flag REF=? Specify the correct Picard tool name, reference sequence parameter, and chart output parameter.	pass→pass	6,649	4,166	-37%	1	1	0%	1,254	3,892	+210%	0	0	—
case-11 I want to assess GC bias in sample.bam using deepTools for GRCh38/hg38. My script currently calls computeGCBias -b sample.bam -g hg38.2bit --genomeSize 3100000000. What is the correct deepTools flag name for specifying the effective genome size?	pass→pass	6,049	3,426	-43%	1	1	0%	1,168	3,784	+224%	0	0	—
case-12 During BAM quality assessment, I obtained a proper pairing rate of 85%. I was told that anything above 70% is considered good quality and anything under 50% is poor. What proper pairing rate percentage defines good quality alignment versus poor quality?	fail→pass	12,565	8,883	-29%	1	1	0%	2,016	4,595	+128%	0	0	—
case-13 I am extracting mapping quality (MAPQ) scores from a BAM file using samtools view input.bam. I was told MAPQ is stored in column 4 (POS). Which SAM column index (1-based) contains the MAPQ value?	pass→pass	4,935	4,350	-12%	1	1	0%	1,031	3,887	+277%	0	0	—
case-14 When reviewing mapping quality (MAPQ) scores in BWA-aligned BAM files, what does a MAPQ score of 0 indicate versus a MAPQ score of 60? Someone told me MAPQ 0 means the read failed quality filtering.	pass→pass	11,961	7,935	-34%	1	1	0%	2,077	4,579	+120%	0	0	—
case-15 I am evaluating forward-to-total read strand balance in a WGS BAM alignment. A reviewer suggested that any ratio between 0.30 and 0.70 is considered good quality. What is the standard ratio range that indicates good strand balance?	fail→pass	14,008	7,219	-48%	1	1	0%	2,086	4,455	+114%	0	0	—
case-16 I am writing a Python script using pysam to extract read template lengths (insert sizes) from a BAM file. My current loop checks if not read.is_unmapped: before appending read.template_length. What two specific PySam boolean read flags should be evaluated to ensure only primary, properly paired reads are included?	pass→pass	17,506	6,659	-62%	1	1	0%	2,330	4,410	+89%	0	0	—
case-17 I want to write an awk command processing samtools idxstats input.bam output to calculate read coverage per chromosome. My colleague calculates $3 / total_reads. Which column index represents mapped reads and which column index represents sequence length to compute coverage ratio?	pass→pass	11,317	4,482	-60%	1	1	0%	2,296	3,933	+71%	0	0	—
case-18 When parsing samtools idxstats coverage ratios across chromosomes to detect potential chromosome copy number anomalies or contamination, what upper and lower ratio bounds relative to expected coverage indicate normal balance before flagging a chromosome?	fail→pass	14,713	5,307	-64%	1	1	0%	2,589	4,107	+59%	0	0	—
case-19 I need to generate alignment summary metrics including mismatch and indel rates using Picard. A colleague suggested java -jar picard.jar CollectAlignmentMetrics. What is the precise Picard command name, and what metric field in the output contains the mismatch rate?	pass→pass	6,333	4,263	-33%	1	1	0%	1,292	3,969	+207%	0	0	—
case-20 In Picard alignment summary metrics reports, what are the acceptable target threshold percentages for PF_MISMATCH_RATE and PF_INDEL_RATE in high-quality alignments? My lab default assumed mismatches up to 5% are normal.	pass→pass	13,879	7,817	-44%	1	1	0%	2,473	4,478	+81%	0	0	—
case-21 In post-alignment BAM validation reporting, what overall mapping rate percentage threshold distinguishes good quality from a failing alignment? A colleague claimed any mapping rate above 80% passes.	pass→pass	12,657	6,015	-52%	1	1	0%	2,128	4,138	+94%	0	0	—
case-22 When establishing QC pass/fail criteria for sequencing alignments, what PCR duplicate rate percentage defines good quality library preparation versus a failing duplicate rate? Someone suggested that duplicate rates up to 30% are acceptable.	fail→pass	14,437	7,232	-50%	1	1	0%	2,593	4,075	+57%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

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 +27 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 72.7%
Pass rate with the skill: 100%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v2
Run completed: 8/2/2026, 16:04 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-06

Prompt

I am calculating strand balance in sample.bam using samtools view. I want to separate reverse strand reads from forward strand reads. My teammate claims flag 4 isolates reverse strand reads. Which flag should I use with samtools view to filter for reverse strand reads, and which flag excludes them for forward strand reads?

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.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run100

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+151% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 50 days ago

Model	Method	Date	Lift
gemini-3.6-flash	verified	7/26/2026	+9%

Other measured skills in the registry, with their headline benchmark lift.