What is the difference between a wet lab and a dry lab?

A wet lab handles chemicals, biological materials and liquids, and needs infrastructure like plumbing and ventilation. A dry lab does computational or analytical work with no physical handling of samples. The wet lab generates the data, the dry lab analyzes it.

What is a dry lab in bioinformatics?

It is the computational side of research: bioinformatics, computational biology, data science and molecular modeling. Work happens on a computer, not at a bench. The dry lab scales well, but it needs wet lab validation to confirm that results reflect real biology.

How do I transition from wet lab to bioinformatics?

Take ownership of one computational task inside your current team, such as QC reports for RNA-seq or differential expression in R, then expand from there. The book's author reports about a year to get comfortable with Unix and two to three years to master a language like R. Your biology background shortens the bridge.

Do bioinformaticians need wet lab experience?

It is not required, but it is an edge. People who have run PCR, cell culture and cloning recognize when a result lacks biological sense and ask better questions. If you lack it, compensate by spending time at the bench with the people who generate your data.

Glossary · Learning and Getting Started

Wet lab vs dry lab

The bench-versus-computer split is where most broken analyses begin, because the data you receive carries every decision made before it reached you.

By Ming "Tommy" Tang, Director of Bioinformatics in Big Pharma · Reviewed October 2026 · 2 min read

Also: dry lab, wet lab

Definition

A wet lab is a laboratory built to handle chemicals, biological materials and liquids, with the plumbing, ventilation and chemical-resistant surfaces that requires. A dry lab is a research space for computational, analytical or data-driven work that does not touch chemicals or biological material: bioinformatics, computational biology, data science, molecular modeling. The wet lab produces the measurements; the dry lab interprets them. Think of the wet lab as where the data is made and the dry lab as where it is read.

You will meet this split the day you become the person who gets sent the data. Someone at the bench ran the experiment, someone at a computer analyzes it, and the two rarely sit in the same room. If you came from the bench, you already know half of what the dry lab needs. If you came from computing, you are missing that half.

The term decides how you handle every odd result. An outlier in a PCA plot, a sample that will not cluster, a low quality metric: each is either a code problem or a lab problem, and you cannot tell which without knowing how the sample was made.

Why it matters

Data carries artifacts from sample preparation, and no algorithm removes an artifact it cannot recognize. Real datasets contain mislabeled samples, batch effects and handling problems, and the only way to find their source is to understand how the data was generated.

A concrete case: a ChIP-seq sample failed to cluster with its group and looked like junk. The cause was chromatin degraded after immunoprecipitation, which only the bench could explain. An ATAC-seq sample with a low TSS enrichment score made biological sense once the experimentalist described the protocol details. Without that conversation, you either drop good data or keep bad data.

The dry lab also cannot validate itself. It scales well and can explore large design spaces, but it cannot tell you whether a result reflects real biology without wet lab feedback.

Where people get it wrong

The usual mistake is treating the two as sequential stages: the wet lab finishes, throws data over the wall, and the dry lab starts. That model makes you analyze data you do not understand. Effective teams define experimental steps together, before data collection, not after.

A second mistake is assuming a sample name means the same thing everywhere. MCF7 in one lab is not MCF7 in another: density, passage number and media change the biology. Computational patterns alone will not tell you that. A third is treating the spreadsheet as a dry lab problem. Messy sample sheets are a wet lab data quality problem that you inherit.

A concrete example

A sample sheet arrives from the bench with a column called "Sample ID #", sex entered as F, Female and female, and a gene symbol that Excel has turned into a date. Before any analysis, check that the sheet is rectangular (one row per sample, one column per variable, no empty or merged cells) and that categories are consistent. Then ask the experimentalist about any sample that looks off, in person if you can.

library(readr)
library(dplyr)

meta <- read_csv("sample_sheet.csv")

# column names should be snake_case with no spaces or symbols
names(meta)

# inconsistent categories show up as extra levels
count(meta, sex)

# empty cells break rectangular data
colSums(is.na(meta))

Related terms

Questions people ask

What is the difference between a wet lab and a dry lab?: A wet lab handles chemicals, biological materials and liquids, and needs infrastructure like plumbing and ventilation. A dry lab does computational or analytical work with no physical handling of samples. The wet lab generates the data, the dry lab analyzes it.
What is a dry lab in bioinformatics?: It is the computational side of research: bioinformatics, computational biology, data science and molecular modeling. Work happens on a computer, not at a bench. The dry lab scales well, but it needs wet lab validation to confirm that results reflect real biology.
How do I transition from wet lab to bioinformatics?: Take ownership of one computational task inside your current team, such as QC reports for RNA-seq or differential expression in R, then expand from there. The book's author reports about a year to get comfortable with Unix and two to three years to master a language like R. Your biology background shortens the bridge.
Do bioinformaticians need wet lab experience?: It is not required, but it is an edge. People who have run PCR, cell culture and cloning recognize when a result lacks biological sense and ask better questions. If you lack it, compensate by spending time at the bench with the people who generate your data.

Guide · How to Detect Batch Effects in Single-Cell RNA-seq
Guide · How to Call Peaks You Can Trust in ChIP-seq
Guide · How to Avoid Pseudoreplication in Single-Cell RNA-seq
Guide · How to Catch a Sample Swap in Variant Calling
Compare · CUT&RUN and CUT&Tag vs ChIP-seq: Which One Should You Use?

Sources

Has AI Changed the Course of Drug Development? Three Years Later — Data quality as the bottleneck; wet lab validation remains essential