sequence-analysis
Workspace gateway for day-to-day sequence work. The current bio environment has real local executables for blastn, blastp, makeblastdb, bowtie2, bwa, samtools, hmmscan, hmmbuild, mafft, muscle, hisat2, featureCounts, prodigal, RNAfold, seqkit, and Biopython.
Quick Start
- Activate environment:
conda activate bio
- Verified local tools:
blastn, blastp, makeblastdb, bowtie2, bwa, samtools, hmmscan, hmmbuild, mafft, muscle, hisat2, featureCounts, prodigal, RNAfold, seqkit
- Verified Python library:
Bio (Biopython)
When To Use This Tool
- Choosing the right installed tool for DNA/RNA/protein sequence analysis
- Running alignment, homology search, ORF prediction, or basic format/statistics work
- Building a small ad hoc sequence workflow without committing to a bigger pipeline
- Verifying what is truly available in the workspace before proposing a method
Common Patterns
# Local BLAST search
blastn -query input.fa -db nt -outfmt 6 -out results.tsv
# Multiple-sequence alignment
mafft input.fa > aligned.fa
# ORF prediction for a microbial genome
prodigal -i genome.fa -a proteins.faa -d genes.fna -f gff -o genes.gff
# Basic Biopython parsing
from Bio import SeqIO
records = list(SeqIO.parse("input.fa", "fasta"))
print(len(records))
Recommended Workflow
- Activate the
bio environment and confirm the exact tool you need is present.
- Validate the input format before launching expensive searches or alignments.
- Prefer the simplest installed CLI that matches the task: BLAST for homology search, MAFFT/MUSCLE for MSA, HMMER for domain search, Prodigal for microbial ORFs, and Samtools/Hisat2/FeatureCounts for RNA-seq-style alignment/counting.
- Escalate to more specialized project skills only when the plain CLI workflow is clearly insufficient.
Guardrails
- Keep claims limited to tools verified on
PATH in this workspace.
- The old autogenerated skill suggested direct Biomni tool imports as a ready path, but local Biomni tool imports currently fail because
langchain_core is missing.
- Use CLI-first workflows here; treat Biomni/Evo2 as separate repo-backed projects with extra dependency requirements.
- Validate input formats explicitly. Many of these tools assume FASTA/FASTQ/SAM/BAM conventions and will fail noisily or misleadingly on malformed input.
1---2name: sequence-analysis3description: Use when routing DNA, RNA, or protein sequence tasks to the core sequence-analysis commands that are actually installed in this workspace.4---5
6# sequence-analysis
7
8Workspace gateway for day-to-day sequence work. The current `bio` environment has real local executables for `blastn`, `blastp`, `makeblastdb`, `bowtie2`, `bwa`, `samtools`, `hmmscan`, `hmmbuild`, `mafft`, `muscle`, `hisat2`, `featureCounts`, `prodigal`, `RNAfold`, `seqkit`, and Biopython.
9
10## Quick Start
11
12- **Activate environment:** `conda activate bio`
13- **Verified local tools:** `blastn`, `blastp`, `makeblastdb`, `bowtie2`, `bwa`, `samtools`, `hmmscan`, `hmmbuild`, `mafft`, `muscle`, `hisat2`, `featureCounts`, `prodigal`, `RNAfold`, `seqkit`
14- **Verified Python library:** `Bio` (Biopython)
15
16## When To Use This Tool
17
18- Choosing the right installed tool for DNA/RNA/protein sequence analysis
19- Running alignment, homology search, ORF prediction, or basic format/statistics work
20- Building a small ad hoc sequence workflow without committing to a bigger pipeline
21- Verifying what is truly available in the workspace before proposing a method
22
23## Common Patterns
24
25```bash
26# Local BLAST search
27blastn -query input.fa -db nt -outfmt 6 -out results.tsv
28```
29
30```bash
31# Multiple-sequence alignment
32mafft input.fa > aligned.fa
33```
34
35```bash
36# ORF prediction for a microbial genome
37prodigal -i genome.fa -a proteins.faa -d genes.fna -f gff -o genes.gff
38```
39
40```python
41# Basic Biopython parsing
42from Bio import SeqIO
43records = list(SeqIO.parse("input.fa", "fasta"))
44print(len(records))
45```
46
47## Recommended Workflow
48
491. Activate the `bio` environment and confirm the exact tool you need is present.
502. Validate the input format before launching expensive searches or alignments.
513. Prefer the simplest installed CLI that matches the task: BLAST for homology search, MAFFT/MUSCLE for MSA, HMMER for domain search, Prodigal for microbial ORFs, and Samtools/Hisat2/FeatureCounts for RNA-seq-style alignment/counting.
524. Escalate to more specialized project skills only when the plain CLI workflow is clearly insufficient.
53
54## Guardrails
55
56- Keep claims limited to tools verified on `PATH` in this workspace.
57- The old autogenerated skill suggested direct Biomni tool imports as a ready path, but local Biomni tool imports currently fail because `langchain_core` is missing.
58- Use CLI-first workflows here; treat Biomni/Evo2 as separate repo-backed projects with extra dependency requirements.
59- Validate input formats explicitly. Many of these tools assume FASTA/FASTQ/SAM/BAM conventions and will fail noisily or misleadingly on malformed input.
vimalinx