LabGenius

LabGenius is a next-gen lab inventory web-app powered by AI.

Comment

LabGenius – The Genesis of Great Ideas

Demo | GitHub | See our presentation

Putting scientists' needs first

LabGenius is here to simplify the scientific processes for scientists by eliminating the need to keep a mental map of all your experimental materials and lowering the barrier for intra-laboratory collaborations. LabGenius keeps track of the locations of plasmids, proteins, and cell lines in your freezer, fridges, and benches. LabGenius also lets you contextualize your entire lab's materials with Collaborative Search™.

Next-generation Lab Inventory

LabGenius stores files, volumes, concentrations, locations, and notes on all your lab materials. Enter general information and/or files (.fasta, .faa, .gbk, .dna, etc.) for plasmids, proteins, or cell lines. LabGenius will store them and help you keep track of them!

Example:

You: I have put away 100 µL of pET-24a-GFP-6xHIS at 135 ng/mL in my Main Project box

LabGenius: Great! I have stored that away for you. Would you like to provide a specific box location or file?

You: A10

You: `pET-24a-GFP-6xHIS.gbk` Uploaded

LabGenius: I have detected primers in your plasmid. I have put those away, too! You can provide more information on primers later.

Collaborative Search™

LabGenius lets you contextually search lab materials you or other group members have generated with Collaborative Search™. Ask the Genius for a record of specific samples or ask if there is something available you need that your lab might have!

Example:

You: Do I have any anti-FLAG antibodies?

LabGenius: Hmm, let me check that for you...

LabGenius: You do not have anti-FLAG antibodies, but it looks like someone else in your lab does. Would you like a list of your available antibodies or where to find anti-FLAG?

You: Where can I find anti-FLAG?

LabGenius: It looks like Daniel has anti-FLAG aliquots (20 µL) in his Communal Freezer but has not specified a location. Hope this helps!

Let us show you what we're made of

Great tools run on great platforms! LabGenius is powered by Vertex AI & Google PaLM 2 all within Google Cloud.

At our cores...

We have taken a security first and scalability second approach when storing your scientific data. That is why we have chosen to host all data in Firebase and our compute on Google Cloud.

It takes one to know one...

The genius behind LabGenius is Vertex AI which is powered by Google's PaLM 2. With a touch of fine-tuning, Vertex AI is turned into Collaborative Search™ & able to access your lab's materials contextually.

Challenges in our fields

Nothing worth doing is ever easy. We recognize that idealizing technology only leads to dead ends. That is why we have identified our most foreseeable pain points.

Embedding complex scientific data

Laboratories create a wide array of data. Some of it can be stored on computers, but most is stored in freezers, fridges, and benches. Although DNA and Proteins have digitized sequences associated, their scientific contexts are equally valuable information that isn't so easily compiled into 1's and 0's. LabGenius' vector database currently embeds all data using the same algorithm. It is easy to see the downfall of this method. Although streamlined, it can lead to suboptimal results when building contextual relationships within data types.

Training & fine-tuning AI models for science

LabGenius thrives to satisfy a wide gamut of needs. As we all know, not all scientific fields have the exact needs or data types that require inventorying. This has made training and fine-tuning Collaborative Search™ a gargantuan task with current tools. We are far from perfect, and the foundation of LabGenius likely will require re-building (although not from the ground up).

Maintaining good lab practices

The most challenging aspect of being a scientist is developing great ideas. The second is documenting your discoveries. LabGenius works best with constantly updated information. This makes our biggest challenge getting graduate students and scientists accustomed to recording data they usually just keep in their heads. We want you to speak your mind!

What's in our future?

Prediction is very difficult, especially if it's about the future. –Nils Bohr

Although we don't want to be overzealous & say what the future has in store for us, we do have some long-term goals.

Expanding the data we can handle

We plan on incorporating unstructured input with the help of our users. LabGenius is not just an inventory app; it's an AI that is good at keeping track of lab materials. This means LabGenius can learn and grow with the labs that use it. Incorporating a "learn" mode is a milestone we hope to eventually achieve.

Simplifying input & interactions

The challenge we are most excited about tackling is Maintaining good lab practices. We plan to address this by allowing scientists to interact with LabGenius multi-modally via text, mobile app, or web. Our first order of business is implementing Text-to-Genius™ to allow you to interact with LabGenius via text messages powered by Twilio.

Built With

Share this project:

Updates