Introduction

The philosophy of DSH (DeepSeek Harness) is “everything is a plugin”. When adding domain capabilities to an agent, the most common burden is dependencies: calling external command-line tools requires a pre-installed runtime environment, while calling online services is constrained by network and third-party availability. However, much of the computation in molecular biology—sequence transformation, primer Tm, restriction enzyme simulation—can be done locally deterministically without needing to carry a string of dependencies.

dsh-molbio-tools is a plugin implemented according to this philosophy: a zero-dependency molecular biology research plugin for DeepSeek Harness, covering sequence analysis, primer auto-design and checking, restriction enzyme simulation, GenBank parsing and plasmid maps, qPCR analysis, literature assistant, and benchtop calculations. Below is an introduction in order of positioning, features, installation, and typical usage.

What is it

dsh-molbio-tools (repository lovy2004/dsh-molbio-tools, package name dsh-molbio-tools) is maintained by lovy2004 and released under the MIT license. It has three core characteristics:

  1. 44 tools in total, all composed of zero-dependency pure .mjs files, which can be distributed together with the agent preset directory;
  2. Except for literature retrieval and storage, everything is deterministic pure computation;
  3. Installed in a dedicated preset mode, where molbio tools only appear in the corresponding session mode.

It solves the problem of: putting the workflow of sequence analysis, primer design, cloning simulation and verification, qPCR analysis, literature and experiment records into a single DSH session, requiring no installation of any external programs throughout the process.

Core Features

Sequence Analysis and Restriction Enzymes

  • molbio_reverse_complement: Reverse complement / Complementary strand, supports IUPAC degenerate codes
  • molbio_gc_content: Overall GC content, optional windowing
  • molbio_translate: 1/3/6 frame translation and ORF finding
  • molbio_restriction_sites: Built-in search for restriction enzyme sites of 90+ enzymes and calculation of enzyme fragments (since v10 includes Type IIS enzymes)
  • molbio_enzyme_lookup: v13 enzyme directory query; reports double-stranded cut points for a given sequence, including Type IIS enzyme cuts where the reverse recognition site is cut (a case not reported by molbio_restriction_sites)

Primer Design and Checking

  • molbio_design_primers: Automatic design of PCR primer pairs; Tm uses the SantaLucia 1998 NN model; since v12, Primer3-style structural screening and mismatch tolerance; since v13, salt/concentration knobs and 3’ target position preference
  • molbio_design_intron_primers: v10 intron-spanning qPCR primers, forward primer spans exon-exon junction, fails to amplify when using genomic DNA as template
  • molbio_primer_tm: Single primer Tm estimation
  • molbio_primer_check: Primer structural screening (repeats, self-complementarity, hairpins, dimers)

Plasmid Maps and Cloning Construction

  • molbio_parse_snapgene / molbio_parse_genbank: SnapGene .dna and GenBank flatfile parsing
  • molbio_plasmid_map / molbio_plasmid_map_file: SVG plasmid map rendering (circular/linear, feature tracks, GC skew, single enzyme marks), writes directly to workspace files and opens automatically
  • molbio_unique_cutters: Unique cutter enzymes
  • molbio_clone_simulate: Restriction enzyme digestion and ligation and Gibson assembly simulation, outputting feature coordinate shifts, verification of digestion predictions, and FASTA/SVG
  • molbio_golden_gate: v13 Golden Gate multi-fragment assembly based on Type IIS enzymes with automatic overhang design
  • molbio_clone_primers: Primer tailing for cloning
  • molbio_mutagenesis_primers: QuickChange-style site-directed mutagenesis primers

Cloning Verification, Experimental Statistics, and Proteins

  • molbio_verify_sanger: Reads .ab1/.seq etc. sequencing files, circular-aware alignment against reference plasmid, reports mismatches/deletions/insertions, identity, and consequences for CDS amino acids
  • molbio_qpcr_analysis: qPCR ΔΔCt analysis
  • molbio_qpcr_efficiency: Dilution series amplification efficiency standard curves (SVG)
  • molbio_plot / molbio_virtual_gel: General SVG bar/scatter plots; v13 virtual agarose gel
  • molbio_lab_math: Dilution, molarity, DNA copy number calculations
  • molbio_protein_props: MW / pI / A280 / GRAVY / hydrophobicity index estimation
  • molbio_peptide_digest: Peptide digestion simulation for mass spectrometry
  • molbio_codon_optimize: E. coli/Yeast/Human codon optimization, can avoid specified restriction sites

Sequence Extension and Multiple Sequence Alignment

  • molbio_align: Smith-Waterman local alignment
  • molbio_fasta_fastq: FASTA/FASTQ processing and QC
  • molbio_extract_region: Extract subsequence by feature name or coordinates
  • molbio_msa_align (v15): Progressive multiple sequence alignment (affine gap NW + 5-mer/UPGMA guide tree)
  • molbio_conservation (v15): Conservation analysis, outputting consensus sequence, per-column identity, and entropy scores

Literature and Experimental Records

  • molbio_pubmed_search / molbio_pubmed_abstract: PubMed search and abstract retrieval by PMID
  • molbio_paper_add / molbio_paper_list / molbio_paper_update / molbio_paper_remove: CRUD and deduplication for the literature reading library
  • molbio_paper_export_bibtex: Export library to BibTeX
  • molbio_protocol_add / protocol_list / protocol_update: Protocol library
  • molbio_experiment_log / experiment_list: Experiment logs

Installation and Activation

The recommended installation method is to copy the complete preset directory rather than registering tools one by one. The repository’s preset/molbio-lab/ is the complete preset directory. Copy it to the harness user directory and execute in the repository root:

cp -r preset/molbio-lab ~/.dsh/.agent-presets/molbio-lab

Directory structure after copying:

~/.dsh/.agent-presets/molbio-lab/
├── agent.cordis.yml
├── preset.yml
└── plugins/
    └── dsh-molbio-tools-v15/

After restarting or refreshing the preset list, select “Molecular Biology Lab” in the preset selector to create a new session. After installation, molbio tools only appear in that mode and will not inject the 44 tools and prompt text into other sessions.

Note one rule when updating the plugin: DSH’s standing mount caches modules by ESM module URL, so a new version directory must be created every time an update occurs (e.g., changing dsh-molbio-tools-v15/ to a new version number directory).

Typical Usage

One-Step Plasmid Map Generation

First prepare a .dna (SnapGene) or .gb/.gbk (GenBank) file, then call molbio_plasmid_map_file: The tool reads the file directly, writes the SVG plot to the workspace and returns svg_path (default <name>.svg, can be specified with output_path), and opens it automatically with the system default application after writing (auto_view is enabled by default).

Primer Design and Non-Specific Binding Check

Use molbio_design_primers to automatically design primer pairs. When check_mispriming: true is set, the tool checks for non-specific binding sites of the 3’ ends on the template and rejects or penalizes them; for SNP / site-directed design, since v13, you can use target_position + target_penalty to sort by “distance from the primer 3’ end to the target site”, reporting target_distance for each primer pair.

qPCR Standard Curve

molbio_qpcr_efficiency uses a dilution series to generate a standard curve, plot_path writes the standard curve with fitting lines directly as an SVG and opens it automatically.

From Multiple Sequence Alignment to Conservation Analysis

First call molbio_msa_align to perform multiple sequence alignment, returning the aligned sequences (an array of sequences or fasta_path), and save_path can directly write out the aligned FASTA; then pass that output to molbio_conservation to obtain the consensus sequence, per-column identity, and entropy scores.

Scenarios and Notes

Suitable for two types of users: molecular biologists who perform sequence analysis, primer design, cloning construction and verification within DSH sessions; and DSH preset developers who want to reference the organization method of zero-dependency plugins.

Note the following points before use:

  1. molbio_pubmed_search relies on the harness’s web search service, molbio_pubmed_abstract relies on the deployment providing web fetch capability; it will report an error clearly if not supported;
  2. The literature library storage is subject to harness fs service and sandbox policies;
  3. The numerical values in molbio_protein_props are all estimates; codon optimization is based on published high-frequency codon tables (heuristic);
  4. The plugin runs with the current dsh process permissions; you should check the source code and license (MIT) before installing.

Summary

dsh-molbio-tools packs a molecular biology workflow into 44 zero-dependency pure computation tools; installation is just copying a preset directory, and after enabling, it only affects the dedicated session. For teams needing to do sequence and cloning work in DSH, this is a ready-to-copy starting point.

Related Links:

  • Plugin Directory Page: https://www.skillhub.cn/plugins/lovy2004/dsh-molbio-tools
  • GitHub Repository: https://github.com/lovy2004/dsh-molbio-tools

The community plugin directory is an independent site with no official affiliation to DeepSeek or Hongfang.