Use Cases

A platform so flexible you can run any data, any analysis, for any type of science.

Library Screens

Library based, high throughput drug design

Many biologic therapeutic modalities are easy to synthesize in combinatorial libraries, including antibodies, RNA binding therapies, viral vectors, and more. Using these methods with high-throughput library screening, biotechs are able to evaluate the fitness of hundreds of thousands of potential drug candidates against different targets or other types of selection.

diagram showing the steps of the this use case

Manage NGS experiments

Biotechs often use sequencing to count performant library members after an enrichment assay, resulting in large FASTQ files. These files are difficult to share, move around, and explore.

Here a bioinformatician populates Latch Data. They may upload files from AWS s3, BaseSpace, FTP, SRA, or local computers. They may do this from the comfort of their terminal. is a cloud native filesystem that can store and version all of your sequencing files.

A biologist can then easily access, QC, organize, and explore their sequencing data. Scientists do not need anyone else to do this for them after it has uploaded.

screenshot closeup of latch data filesscreenshot closeup of latch data files

Identify Enriched Library Members

Pooled library screens allow for hundreds, or even thousands, of candidate genes to be screened simultaneously. A bulk measurement is then taken from the pooled sequencing product, and computationally untangled to reveal which candidates are enriched or depleted in the treatment group.

Here a bioinformatician can use the Latch SDK to create production-grade QC and barcode counting workflows to process library members from raw sequencing files. They can easily iterate on the parameters and errors with biologists with full control over code.

screenshot of an example panning enrichment workflow parameter interface

Here as a biologist, thank your bioinformatician. You now have a simple interface to analyze your enrichment data yourself.


Compare Library Performance across Conditions

When working with large-scale library screens under different selection conditions, it can be challenging to choose a “winner” given experimental variability. Statistical analysis is used to guide further investigation by narrowing in on lead candidates.

screenshot of an example comparison graphs

A bioinformatician can use Latch Pods to write flexible, custom jupyter or Rstudio notebooks to plot enrichment counts. Build an application to compile your results and expose it directly to your scientists without worrying about hosting or infrastructure.

The biologist can then easily interact with plots, allowing them to change filters and parameters on custom-built applications. For example, an application may display top performing library members from multiple enrichment experiments.


Organize Enrichment Data for Longitudinal Analysis

When evaluating libraries of candidate therapeutics against numerous conditions, the complexity of the associated sequencing data can quickly become unwieldy. It is critical to clearly register your past experimental data in an organized manner, to enable searching of previous results, contextualization of new data, and synthesis of new conclusions.

Latch Registry is like Notion or Google Sheets but designed for sample sheets. Registry lets you impose structure and organization on files, associating experimental inputs and outputs with metadata.

screenshot of an example registry data organization

A bioinformatician team can build type-validated tables in Registry The bioinformatics are able to enforce accurate data entry, easily clean tables in code, and ingest them into bioinformatics workflows for batch processing.

The biologist can then go in and fill in the samplesheet metadata in a single place. They are easily able to view past experiments in a table that displays both experimental metadata and files in single location.


Metagenomics Classification

Next-generation sequencing, specifically ‘metagenomics’, has enabled a fundamentally new understanding of the microbial world. The gut microbiome has been implicated in a range of health outcomes from obesity, to Alzheimer’s, to immunotherapy response. R&D groups monitoring bacterial species diversity and discovering microbiome-derived therapeutic candidates need access to large-scale data analysis capabilities from day one.

diagram showing the steps of the this use case

Manage all metagenomics data

The monitoring of microbial samples over time generates large collections of FastQ files that will only grow in number. Storage and processing infrastructure must flex accordingly.

screenshot closeup of latch data filesscreenshot closeup of a latch registry table

Latch allow your team monitor your microbial samples as large collections of FastQ files within latch. Import them to Latch Data and then organize/associate them with metadata in Latch Registry


Obtain Amplicon Sequence Variants (ASVs) to find sample composition

Today’s metagenomics sequencing technologies generate a variety of read lengths and outputs. You must stitch reads together and assemble them into genes or organisms. Complex multi-step bioinformatics pipelines are needed for insights like variant identification.

screenshot closeup of parameters on a dada classification workflow

A bioinformatician can use the Latch SDK to develop a variant identification workflow with a user-friendly interface for other scientists/biologists to run. Or they may use common metagenomics tools which are available on Latch:


Explore phylogeny and taxonomy. Identify microbial features for more research.

Generate visualizations from taxonomic classification data in your browser. Explore alpha or beta diversity and phylogenetic relationships. All of this in custom, templated notebooks.

abundance diversity graphgraph sequence percentage

A bioinformatician can develop a Latch Template that serve Shiny applications like Phylo-seq for scientists to track the dynamic evolution of a microbial population over time. The biologists on their team simply click “Use Template” to begin their own analysis, without having to set up an environment or install dependencies.

Pod Templates on latch

Elegantly Store Data and Organize for Longitudinal Analyses

Take raw outputs from taxonomic classification pipelines, as well as analysis reports, and attach metadata. Organize samples based on characteristics like time point or Patient ID in Latch Registry.

Pod Templates on latch

Query data to assemble comparisons of certain samples based on time point.

Pod Templates on latch


Immunoreceptor Discovery & Screening

Immunomodulation is an increasingly important strategy for combating cancer, infections, and inflammatory diseases. Fine understanding of immune cell phenotypes and T-/B-cell specificities directly translates to the creation of innovative therapeutics, such as mRNA vaccines and engineered T cells. These processes rely on TCR/BCR complex discovery and characterization. Previously a laborious multi-year endeavor, the process of finding these receptors can now be condensed into a matter of weeks using RNAseq.

Let’s explore a common use case enabled by single-cell genomics: paired ɑ/β TCR sequencing.

diagram showing the steps of the this use case

Derive paired ɑ/β TCR nucleotide sequences from sc-RNAseq assays

Single-cell genomics has allowed for the preservation of paired ɑ/β TCR transcripts, which would otherwise be difficult to obtain from bulk samples. Organizing TCR sequences by themselves, or with associated gene expression files, can be complex.

screenshot closeup of latch data files

A bioinformatician can import all this data easily in Latch Data. And then specific TCR + GEX sequences can be organized intuitively for biologist counterparts using Latch Registry.

screenshot closeup of latch data files

After all of the data has been imported, a biologist can easily find TCR + GEX sequences of interest based on functional annotations and clonotype IDs.


Align V(D)J fragments to germline and annotate clonotypes

Immunoreceptor genes are some of the most difficult to sequence on account of the biological complexity of V(D)J recombination. To begin, full-length transcripts must be obtained at the RNA level, then fragmented into V(D)J regions for alignmentment with the germline contig.

To run the analysis, a biologist can drag and drop V(D)J FastQ files, then launch cellranger vdj to process sequence fragments from 10X and generate clonotype annotations — without any code.

screenshot closeup of latch workflow parameters

Your teams bioinformatician can also leverage the Latch SDK to upload your own custom V(D)J alignment workflow from the command line. Any workflow uploaded through the SDK allows biologists to run the workflow in a no-code GUI. It also seamlessly integrate new program updates, letting Latch auto-version your previous work.

screenshot of the workflow development page

Explore TCR clonality and generate visualizations

The immune repertoire contains incredible diversity of clonotype sequences. Observing TCR/BCR clonality over time can generate important insights into disease states and mechanisms of therapeutic response.

A biologist can explore TCR/BCR clonal dynamics using a public Pod Templates like Immunarch. View clone size and expansion kinetics over time. Extract TCR sequences of interest.

screenshot of the workflow development pagescreenshot of the workflow development page

A bioinformatician can generate custom R and Python notebooks for scientist colleagues to run independently. Deliver Jupyter Notebook and R Markdown reports to collaborators using Pod Templates .

Protein Engineering

Accelerate Protein Design

Rational design of proteins can aid a multitude of translational R&D efforts. Models like AlphaFold have been used forin silico SARS-CoV-2 antigen predictions, de novo protein design, and characterizing agents of antibiotic resistance. However usage is blocked by access to GPUs and ease of use for scientists. Latch offers efficient GPU access for running hundreds of protein-ML applications.

diagram showing the steps of the this use case

Launch AlphaFold, ColabFold, or custom models to predict 3D protein structure

Predict 3D protein structures using ready-to-run LLMs.

example parameters on alphafold workflow

Anyone can use the Latch Verified AlphaFold or ColabFold workflow to 3D protein structure from amino acid inputs. Define GPU resources for efficient and rapid LLM inference.


Visualize a .pdb on Latch, or import into a Pod for further investigation

Armed with 3D protein structures, infer functional properties of proteins and explore annotations.

example parameters on alphafold workflow

Or anyone familiar with code can run custom downstream analysis in pods to look at molecular dynamics, binding affinities, and other protein attributes using tools like Bio3D.