Ready to supercharge your AI agents?

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

Bio Pileup Generation · FreedomIntelligence | DecimalAI

Bio Pileup Generation

SkillScore96Maintained

bio-pileup-generation

Generate pileup data for variant calling using samtools mpileup and pysam. Use when preparing data for variant calling, analyzing per-position read data, or calculating allele frequencies.

scienceby FreedomIntelligence

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

Automatic

Coding pack

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

pip install decimalai

2 · Get the skill — writes .claude/skills/bio-pileup-generation/SKILL.md

decimalai skills pull bio-pileup-generation --out .claude/skills/

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

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

Raw file: /s/bio-pileup-generation/SKILL.md ↗ · machine-readable summary at /s/bio-pileup-generation.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

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 49d ago

Yes — +4% lift · at higher cost+172% tokens0% turns96% of whole cases pass with skill (22/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-01	✗→✓	▲ Improved	141%	0%
case-02	✓→✓	＝ Same ✓	298%	0%
case-03	✓→✓	＝ Same ✓	244%	0%
case-04	✓→✓	＝ Same ✓	104%	0%
case-05	✓→✓	＝ Same ✓	152%	0%

Showing 5 of 23 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-pileup-generation/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

-->

Pileup Generation

Generate pileup data for variant calling and position-level analysis.

What is Pileup?

Pileup shows all reads covering each position in the reference, used for:

Variant calling (with bcftools)
Coverage analysis
Allele frequency calculation
SNP/indel detection

samtools mpileup

Basic Pileup

bash
samtools mpileup -f reference.fa input.bam > pileup.txt

Pileup for Variant Calling (Output BCF)

bash
samtools mpileup -f reference.fa -g input.bam -o output.bcf

Pileup Specific Region

bash
samtools mpileup -f reference.fa -r chr1:1000000-2000000 input.bam

Regions from BED

bash
samtools mpileup -f reference.fa -l targets.bed input.bam

Multiple BAM Files

bash
samtools mpileup -f reference.fa sample1.bam sample2.bam sample3.bam > pileup.txt

Output Format

Text pileup format (6 columns per sample):

chr1    1000    A    15    ...............    FFFFFFFFFFF
chr1    1001    T    12    ............      FFFFFFFFFFFF

| Column | Description | |--------|-------------| | 1 | Chromosome | | 2 | Position (1-based) | | 3 | Reference base | | 4 | Read depth | | 5 | Read bases | | 6 | Base qualities |

Read Bases Encoding

| Symbol | Meaning | |--------|---------| | . | Match on forward strand | | , | Match on reverse strand | | ACGT | Mismatch (uppercase = forward) | | acgt | Mismatch (lowercase = reverse) | | ^Q | Start of read (Q = MAPQ as ASCII) | | $ | End of read | | +NNN | Insertion of N bases | | -NNN | Deletion of N bases | | * | Deleted base | | > / < | Reference skip (intron) |

Quality Filtering Options

Minimum Mapping Quality

bash
samtools mpileup -f reference.fa -q 20 input.bam

Minimum Base Quality

bash
samtools mpileup -f reference.fa -Q 20 input.bam

Combined Quality Filters

bash
samtools mpileup -f reference.fa -q 20 -Q 20 input.bam

Maximum Depth

bash
# Prevent memory issues with high coverage
samtools mpileup -f reference.fa -d 1000 input.bam

Variant Calling Pipeline

mpileup to bcftools call

bash
samtools mpileup -f reference.fa input.bam | bcftools call -mv -o variants.vcf

Direct BCF Output

bash
samtools mpileup -f reference.fa -g -o output.bcf input.bam
bcftools call -mv output.bcf -o variants.vcf

Full Pipeline

bash
samtools mpileup -f reference.fa -q 20 -Q 20 input.bam | \
    bcftools call -mv -Oz -o variants.vcf.gz
bcftools index variants.vcf.gz

pysam Python Alternative

Basic Pileup

python
import pysam

with pysam.AlignmentFile('input.bam', 'rb') as bam:
    for pileup_column in bam.pileup('chr1', 1000000, 1001000):
        print(f'{pileup_column.reference_name}:{pileup_column.pos} depth={pileup_column.n}')

Access Reads at Position

python
import pysam

with pysam.AlignmentFile('input.bam', 'rb') as bam:
    for pileup_column in bam.pileup('chr1', 1000000, 1000001, truncate=True):
        print(f'Position: {pileup_column.pos}')
        print(f'Depth: {pileup_column.n}')

        for pileup_read in pileup_column.pileups:
            if pileup_read.is_del:
                print('  Deletion')
            elif pileup_read.is_refskip:
                print('  Reference skip')
            else:
                qpos = pileup_read.query_position
                base = pileup_read.alignment.query_sequence[qpos]
                qual = pileup_read.alignment.query_qualities[qpos]
                print(f'  {base} (Q{qual})')

Count Alleles at Position

python
import pysam
from collections import Counter

def allele_counts(bam_path, chrom, pos):
    counts = Counter()

    with pysam.AlignmentFile(bam_path, 'rb') as bam:
        for pileup_column in bam.pileup(chrom, pos, pos + 1, truncate=True):
            if pileup_column.pos != pos:
                continue

            for pileup_read in pileup_column.pileups:
                if pileup_read.is_del:
                    counts['DEL'] += 1
                elif pileup_read.is_refskip:
                    continue
                else:
                    qpos = pileup_read.query_position
                    base = pileup_read.alignment.query_sequence[qpos]
                    counts[base.upper()] += 1

    return dict(counts)

counts = allele_counts('input.bam', 'chr1', 1000000)
print(counts)  # {'A': 45, 'G': 5}

Calculate Allele Frequency

python
import pysam
from collections import Counter

def allele_frequency(bam_path, chrom, pos, min_qual=20):
    counts = Counter()

    with pysam.AlignmentFile(bam_path, 'rb') as bam:
        for pileup_column in bam.pileup(chrom, pos, pos + 1, truncate=True,
                                         min_base_quality=min_qual):
            if pileup_column.pos != pos:
                continue

            for pileup_read in pileup_column.pileups:
                if pileup_read.is_del or pileup_read.is_refskip:
                    continue
                qpos = pileup_read.query_position
                base = pileup_read.alignment.query_sequence[qpos]
                counts[base.upper()] += 1

    total = sum(counts.values())
    if total == 0:
        return {}

    return {base: count / total for base, count in counts.items()}

freq = allele_frequency('input.bam', 'chr1', 1000000)
for base, f in sorted(freq.items(), key=lambda x: -x[1]):
    print(f'{base}: {f:.1%}')

Pileup with Quality Filtering

python
import pysam

with pysam.AlignmentFile('input.bam', 'rb') as bam:
    for pileup_column in bam.pileup('chr1', 1000000, 1001000,
                                     truncate=True,
                                     min_mapping_quality=20,
                                     min_base_quality=20):
        print(f'{pileup_column.pos}: {pileup_column.n}')

Generate Pileup Text

python
import pysam

def pileup_text(bam_path, ref_path, chrom, start, end):
    with pysam.AlignmentFile(bam_path, 'rb') as bam:
        with pysam.FastaFile(ref_path) as ref:
            for pileup_column in bam.pileup(chrom, start, end, truncate=True):
                pos = pileup_column.pos
                ref_base = ref.fetch(chrom, pos, pos + 1)
                depth = pileup_column.n

                bases = []
                for pileup_read in pileup_column.pileups:
                    if pileup_read.is_del:
                        bases.append('*')
                    elif pileup_read.is_refskip:
                        bases.append('>')
                    else:
                        qpos = pileup_read.query_position
                        base = pileup_read.alignment.query_sequence[qpos]
                        if base.upper() == ref_base.upper():
                            bases.append('.' if not pileup_read.alignment.is_reverse else ',')
                        else:
                            bases.append(base.upper() if not pileup_read.alignment.is_reverse else base.lower())

                print(f'{chrom}\t{pos+1}\t{ref_base}\t{depth}\t{"".join(bases)}')

pileup_text('input.bam', 'reference.fa', 'chr1', 1000000, 1000100)

Pileup Options Summary

| Option | Description | |--------|-------------| | -f FILE | Reference FASTA (required) | | -r REGION | Restrict to region | | -l FILE | BED file of regions | | -q INT | Min mapping quality | | -Q INT | Min base quality | | -d INT | Max depth (default 8000) | | -g | Output BCF format | | -u | Uncompressed BCF output |

Quick Reference

| Task | Command | |------|---------| | Basic pileup | samtools mpileup -f ref.fa in.bam | | Quality filter | samtools mpileup -f ref.fa -q 20 -Q 20 in.bam | | Region | samtools mpileup -f ref.fa -r chr1:1-1000 in.bam | | BCF output | samtools mpileup -f ref.fa -g in.bam -o out.bcf | | To bcftools | samtools mpileup -f ref.fa in.bam \| bcftools call -mv |

Common Errors

| Error | Cause | Solution | |-------|-------|----------| | No FASTA reference | Missing -f option | Add -f reference.fa | | Reference mismatch | Wrong reference | Use same reference as alignment | | Out of memory | High coverage region | Use -d to cap depth |

Related Skills

alignment-filtering - Filter BAM before pileup
reference-operations - Index reference for pileup
bam-statistics - depth command for coverage
variant-calling - bcftools call from pileup

Illustrative Want to see what bio-pileup-generation 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, 16:34 UTC · 23 cases

Lift

+4%

96% vs 91%

Agent turns

1.0 vs 1.0

Tokens

+172%

3,522.87 vs 1,297.13

fail → pass · 1pass → pass · 21fail → fail · 1

Per-case results (23)

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 need a shell command pipeline that reads `aligned_reads.bam` against `ref_genome.fa` to generate pileup data restricted to target intervals in `capture_targets.bed`. Please apply strict filtering for low mapping confidence and poor base quality, then pipe the result into `bcftools` to output a compressed indexed VCF file containing called variants.	fail→pass	8,607	8,000	-7%	1	1	0%	1,817	4,384	+141%	0	0	—
case-02 I have a BAM file named tumor.bam and a reference genome hg38.fa. I want to output basic text pileup data to raw_pileup.txt. A colleague suggested omitting the reference option since BAM headers have sequence dictionaries. Write the shell command to execute this, using the mandatory reference flag.	pass→pass	4,115	2,930	-29%	1	1	0%	825	3,282	+298%	0	0	—
case-03 Write a command to generate pileup data for three BAM files simultaneously: sample1.bam, sample2.bam, and sample3.bam against genome.fasta into combined.pileup. Should each BAM require its own flag like -i, or are they passed as positional arguments?	pass→pass	4,600	2,835	-38%	1	1	0%	940	3,231	+244%	0	0	—
case-04 I want to generate BCF output directly from pileup for variant calling. Some tools use -b for BCF output, but what flag in samtools mpileup specifically outputs binary BCF format?	pass→pass	9,032	4,881	-46%	1	1	0%	1,789	3,646	+104%	0	0	—
case-05 I need to generate pileup data for a specific single region on chromosome 1 from position 1000000 to 2000000 using ref.fa and input.bam. Should I use -l or -r for a region string format chr1:1000000-2000000?	pass→pass	6,317	2,609	-59%	1	1	0%	1,285	3,242	+152%	0	0	—
case-06 In an ultra-deep target amplicon sequencing experiment with 50,000x coverage, samtools mpileup causes high memory usage. I want to limit the maximum read depth per position to 1000 reads. Which option flag controls max depth in samtools mpileup?	pass→pass	3,191	2,339	-27%	1	1	0%	585	3,141	+437%	0	0	—
case-07 When standard text pileup format is generated, each sample produces 6 columns. What information is stored in column 1, column 2, and column 4?	pass→pass	4,244	1,782	-58%	1	1	0%	847	3,016	+256%	0	0	—
case-08 In column 5 of standard text pileup output (the read bases column), you see characters like '.' and ','. How do you interpret a '.' versus a ',' character?	pass→pass	6,554	2,556	-61%	1	1	0%	910	3,158	+247%	0	0	—
case-09 In standard text pileup read base strings, special characters indicate start and end of read alignments. What exact symbol sequence marks the start of a read versus the end of a read?	pass→pass	4,869	2,415	-50%	1	1	0%	828	3,099	+274%	0	0	—
case-10 In text pileup output, how are deletions from the reference distinguished from reference skips (such as introns in RNA-seq alignments)?	pass→pass	10,867	7,322	-33%	1	1	0%	1,895	4,048	+114%	0	0	—
case-11 Write a Python script using a BAM alignment library (like pysam) to open sample.bam and print the chromosome position and total depth for region chr1:1000000-1001000. What attributes on the pileup column object hold position and total read count?	pass→pass	6,787	4,865	-28%	1	1	0%	1,440	3,771	+162%	0	0	—
case-12 By default in pysam, calling `bam.pileup('chr1', 1000, 1005)` returns columns outside 1000-1005 if reads overlap the boundaries. What parameter must be set to restrict output strictly to positions 1000 to 1005?	pass→pass	3,380	2,839	-16%	1	1	0%	626	3,298	+427%	0	0	—
case-13 In Python using pysam, when iterating over pileup reads at a position, what boolean attributes on `pileup_read` tell you if the alignment position is a deletion or an intron reference skip?	pass→pass	4,861	3,340	-31%	1	1	0%	938	3,367	+259%	0	0	—
case-14 Write a Python function using pysam to calculate allele counts at chr1:1000000 in sample.bam, ensuring bases with base quality below 20 are excluded at the pileup level. What argument handles base quality filtering in `bam.pileup()`?	pass→pass	10,226	8,117	-21%	1	1	0%	2,115	4,491	+112%	0	0	—
case-15 I am implementing custom text pileup output in Python using pysam and FastaFile. If a read base matches the reference, how should forward vs reverse strand matches be rendered in standard pileup format?	pass→pass	8,983	5,094	-43%	1	1	0%	1,889	3,759	+99%	0	0	—
case-16 In samtools mpileup, both `-q` and `-Q` accept integer quality values. If I pass `-q 20 -Q 25`, which one filters mapping quality and which filters base quality?	pass→pass	4,502	2,345	-48%	1	1	0%	745	3,138	+321%	0	0	—
case-17 A user runs `samtools mpileup input.bam > pileup.txt` and receives an error indicating reference sequence missing or 'No FASTA reference'. What parameter was omitted and how should it be added?	pass→pass	4,377	2,194	-50%	1	1	0%	777	3,090	+298%	0	0	—
case-18 Write a complete bash command pipeline that generates pileup data from alignment.bam using ref.fa with quality filters (mapping quality >= 20, base quality >= 20), calls variants using bcftools call (multiallelic mode), and outputs compressed variants.vcf.gz followed by indexing.	fail→fail	7,144	6,188	-13%	1	1	0%	1,583	4,069	+157%	0	0	—
case-19 When streaming output from samtools mpileup into downstream tools in BCF format without saving to disk, compression overhead can slow down processing. Which flag produces uncompressed BCF output?	pass→pass	4,283	2,484	-42%	1	1	0%	779	3,172	+307%	0	0	—
case-20 In pysam, when iterating over columns generated by AlignmentFile.pileup(), how do you apply both a minimum mapping quality threshold of 30 and minimum base quality threshold of 25 during column iteration?	pass→pass	11,521	3,872	-66%	1	1	0%	2,296	3,545	+54%	0	0	—
case-21 I have an un-sorted alignment SAM file named raw_alignments.sam. I need to convert it to BAM, sort it by genomic coordinates using 4 threads, and create a coordinate index file (.bai). What samtools commands should I execute?	pass→pass	5,309	4,564	-14%	1	1	0%	1,127	3,513	+212%	0	0	—
case-22 I already generated a VCF file called raw_variants.vcf using bcftools call. Now I want to apply hard filters using bcftools filter to keep only variants with QUAL > 30 and depth (DP) > 10, outputting to filtered.vcf. What exact bcftools command should be used?	pass→pass	10,792	4,340	-60%	1	1	0%	2,039	3,462	+70%	0	0	—
case-23 Before performing read alignment or pileup, I have raw Illumina paired-end FASTQ files (reads1.fastq.gz and reads2.fastq.gz). I need to trim low-quality trailing bases (Phred score < 20) and filter out reads shorter than 35 bp using fastp. What fastp command should I run?	pass→pass	8,654	6,465	-25%	1	1	0%	1,759	4,104	+133%	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. 23 cases were attempted. The headline lift of +4 percentage points is the difference between those two pass rates over the 23 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 91.3%
Pass rate with the skill: 95.7%
Cases attempted: 23
Cases counted toward lift: 23
Cases that regressed: 0
Skill version measured: v2
Run completed: 8/2/2026, 16:34 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-01

Prompt

I need a shell command pipeline that reads `aligned_reads.bam` against `ref_genome.fa` to generate pileup data restricted to target intervals in `capture_targets.bed`. Please apply strict filtering for low mapping confidence and poor base quality, then pipe the result into `bcftools` to output a compressed indexed VCF file containing called variants.

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 run96

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+175% 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 49 days ago

Model	Method	Date	Lift
gemini-3.6-flash	verified	7/26/2026	+18%

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

Ready to supercharge your AI agents?

Bio Pileup Generation

SkillSafety

Source

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

Pileup Generation

What is Pileup?

samtools mpileup

Basic Pileup

Pileup for Variant Calling (Output BCF)

Pileup Specific Region

Regions from BED

Multiple BAM Files

Output Format

Read Bases Encoding

Quality Filtering Options

Minimum Mapping Quality

Minimum Base Quality

Combined Quality Filters

Maximum Depth

Variant Calling Pipeline

mpileup to bcftools call

Direct BCF Output

Full Pipeline

pysam Python Alternative

Basic Pileup

Access Reads at Position

Count Alleles at Position

Calculate Allele Frequency

Pileup with Quality Filtering

Generate Pileup Text

Pileup Options Summary

Quick Reference

Common Errors

Related Skills

How this skill was measured

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-01

Related skills