Custom Bioinformatics Pipeline Development
Bring us the analysis you keep doing by hand — sequencing, variants, structure, vaccines, molecular biology — and we build the pipeline that does it end to end, tested against results you already trust. Scoped fixed price, quoted after a discovery email. No forms; a person reads every message.
What we build
We are bioinformaticians who write production code. Every pipeline below is scoped to your data and your questions — not a template with your logo on it — and delivered as source your team owns and can rerun without us. If a step is done by a well-known published tool, we use that tool rather than reinventing it, and we tell you exactly which version ran.
NGS analysis pipelines
Raw reads to result tables: quality control, trimming, alignment or assembly, quantification, and a summary report your collaborators can read. Bulk and targeted sequencing, RNA-seq, metagenomic and amplicon workflows.
Built around standard tools such as FastQC/fastp, STAR/BWA, featureCountsVariant analysis pipelines
Germline or somatic variant calling, joint filtering, annotation, and a prioritised table with the evidence attached — reproducible from raw data to the final list with one documented command per stage.
Built around standard tools such as GATK/bcftools, VEP/SnpEffStructural bioinformatics pipelines
Structure preparation, docking and screening workflows, and ranked-result reporting. The docking engines are the same AutoDock Vina and GNINA class of tools our own molecular docking page runs.
Built around standard tools such as AutoDock Vina, GNINAReverse vaccinology pipelines
Genome to candidate antigen list: open reading frame calling, subcellular localization, transmembrane topology, antigenicity and epitope screening, then allergen, toxin and conservation filters — with every intermediate table kept for review.
Multi-step screening with documented thresholds at each filterMolecular biology pipelines
Primer and assay design workflows, CRISPR guide-RNA pipelines, cloning and sequence-manipulation scripts, and QC checks. This is the ground our own primer design engine is built on.
Built around standard tools such as Primer3 with SantaLucia thermodynamicsCustom web tools and reporting
Internal calculators, dashboards, and automated report generation that turn pipeline output into something a lab meeting can use. The public tools on this site are examples of what our team ships.
Browser tools, internal web apps, PDF/CSV report generatorsWhy custom beats buying — and when it does not
Off-the-shelf software wins when your data already fits its assumptions. Custom development wins in three situations we see repeatedly — and we will tell you plainly when not to build.
| Situation | Custom build wins when | Buy instead when |
|---|---|---|
| Data format | Unusual sample sheets, local databases, or an instrument that exports something no tool parses | Your data already fits the tool's documented assumptions |
| Reproducibility | The run must repeat on demand: one documented command, pinned tool versions, expected outputs checked automatically | The product's standard run already produces what you need, repeatably |
| Method ownership | A screening cascade or filter logic that is part of your work and cannot be bought as a product | A published tool with a configuration file covers the whole analysis |
The middle case is the one that bites: a colleague's laptop script that nobody else dares to run. If a published tool with a configuration file solves your problem, that is our answer in the discovery reply — you should not pay us to re-implement something that already exists and works. An honest "you don't need this" saves both sides a project.
How we work
-
Discovery email
You email contact@vigyanllm.in with your field, input data, desired output, and timeline. If you have a paper, protocol, or an existing script that almost works, point us at it. We reply with questions and a clear read on whether this is a build, a configuration, or neither.
-
Written spec and fixed quote
Before any paid work, you receive a written specification: inputs, each processing step, the outputs, how validation will be judged, the delivery window, and one fixed price for the whole scope. If the spec changes later, the price changes only by agreement — never silently.
-
Build
We develop against your example data in short reviewable increments. You see progress as running code and result tables, not a demo at the end. Most projects are written in Python and shell with pinned dependencies; if your team already runs a workflow manager, we deliver in that format.
-
Validation
The pipeline is run against a golden dataset — published results, your own previously hand-checked output, or reference data with known answers — and the comparison is part of the deliverable, including any step where it does not match and why.
-
Handover
You receive the source code, the runnable environment, test data with expected outputs, parameter documentation, and a walkthrough session so your team can run and modify it. The code is yours.
Our own 24-step primer design pipeline is a finished example of this shape: a documented, validated pipeline with defined inputs, checks at each stage, and Primer3/SantaLucia thermodynamics at its core. The step-by-step walkthrough is on that page; the benchmark evidence is on our validation page.
What you get
- Versioned source codeFull project history in Git, delivered to your repository or handed over as an archive.
- Runnable environmentPinned dependencies or a container, so the pipeline runs the same way a year from now.
- Test data with expected outputsA small dataset and the exact results it must produce, so anyone can verify a change.
- Parameter documentationWhat each option means, safe ranges, and which parameters were tuned for your data.
- Run log templateTool and database versions recorded per run, for methods sections and for debugging.
- Walkthrough sessionA live handover call where your team runs the pipeline and asks questions.
How we validate
Validation means the pipeline reproduces results that are already known to be right. Depending on your field, the golden dataset is a published dataset with an expected answer, output you have already checked by hand, or reference material from the same instrument. We run it through, compare stage by stage, and hand you the comparison.
Two honest limits, stated up front. Reproducing a known result shows the pipeline implements the method correctly; it does not guarantee every future dataset will behave identically — input quality still matters, and the pipeline reports what it cannot process instead of guessing. And everything here is in silico: results are research-use-only predictions that your experiments confirm, not diagnostic calls. This matches how we document evidence on our validation page.
Who this is for
Academic labs and CGHS-funded research groups that have outgrown spreadsheets and one-off scripts; CROs standardising a repeatable analysis for clients; vaccine, agricultural and biotech R&D teams with a screening or annotation workflow that runs too often to stay manual; PhD groups that need a defensible, documented method section rather than a folder of notebooks. If you are unsure whether your project fits, ask — the discovery reply is free.
Engagement is deliberately simple: one discovery email, one written spec with a scoped fixed price, then build, validation, and handover. We quote per project because scope drives cost; that is also why we do not publish generic price cards for custom work. Team subscriptions for the self-service tools on this site are separate and listed on /pricing. Data-handling terms are agreed in writing before any file changes hands.
Frequently asked questions
What do you need from me to estimate a project?
Your field of study, what the input data looks like (type and format), what output you need (tables, variants, ranked lists, reports), and your timeline. If a paper, protocol, or existing script shows what you mean, reference it — that usually removes several email round trips.
How much does a custom pipeline cost?
It is quoted as a scoped fixed price after the discovery email, based on the written specification — not billed by the hour. The quote, the scope, and the delivery window are in the same document, and changes to scope are priced only by mutual agreement before the work happens.
How long does a project take?
A focused single-purpose pipeline is typically a few weeks; a multi-step production workflow with deeper validation typically takes a couple of months. Both are indicative ranges, not guarantees — your written spec states the delivery window we commit to for your scope.
Who owns the code you write?
You do. The deliverable is source code handed to your team, with permission to run, modify, and reuse it internally. We do not hold your pipeline hostage behind a licence, and there is no per-run fee after handover.
Can you work with our existing scripts or HPC?
Yes. Common starting points are a script only one person understands, a workflow that runs on a shared cluster, or analysis spread across several tools. We can modernise what exists, integrate it into one pipeline, and shape the delivery for the environment where it will actually run.
What data do I have to send, and where does it go?
For estimation, none — example structures or a small sample of reads are usually enough once work starts. Data-handling terms are agreed in writing before any file changes hands, and your first email should describe data in words rather than attaching it. Never send patient-identifiable data to a first contact.
How do you prove the pipeline works before I rely on it?
By running it on a golden dataset whose correct output is already known — a published dataset or your own hand-checked results — and including that stage-by-stage comparison in the deliverable, along with any step where the outputs differ and why.
Do you also do wet-lab or clinical services?
No. We build in silico analysis pipelines and software. Outputs are research-use-only predictions that your experiments confirm; we do not perform experiments, and nothing delivered is a diagnostic or clinical result.
Start with one email
No form, no sales funnel — a person reads every message sent to contact@vigyanllm.in and replies with questions, an honest read on scope, and then a written spec with a fixed price.
- Field of study and what the project has to produce
- Input data — type and format, described in words
- Desired output: tables, VCF, ranked lists, plots, reports
- Timeline or deadline, if there is one
- A paper, protocol, or existing script to look at first
Custom development is a separate engagement from the self-service tools on this site. Tool plans and limits are always on /pricing.