> For the complete documentation index, see [llms.txt](https://docs.hits.ai/hyperlab-release-note-en/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.hits.ai/hyperlab-release-note-en/changelog_en/2026-09-08-release-note.md).

# 2026-09-08 Release Note

The update for HyperLab on Sep 8, 2026, has been completed.

**HyperLab Version : 2026.09.08**

***

This update introduces molecular structure input and editing with the Ketcher molecular canvas, protein–ligand binding affinity prediction, and new collapse and close controls for question cards.

Researchers can now draw molecular structures directly or modify existing structures and use them as inputs for new research requests. Protein–ligand binding affinity can also be predicted to help compare candidate compounds. In addition, when reviewing questions from the Agent, users can more easily refer back to previous conversations or skip the current question and continue with another research request.

***

## 📢 Small Molecule Discovery Changelogs

### 1. Expanded Chemical Space in Hyper Screening X

The searchable chemical space in **Hyper Screening X** within the Small Molecule Discovery Labspace has been significantly expanded to approximately **200 trillion compounds**.

The building block pool has been substantially expanded, and the maximum number of building blocks that can be connected within a single synthetic route has increased to four.

New synthetic routes have also been added that allow functional groups such as **ester linkers, free amines, and free carboxylic acids** to remain in the final structure.

This expansion enables researchers to explore a much broader chemical space and discover candidate compounds with a wider variety of scaffolds.

***

## 📢 Bio Co-Scientist Changelogs

### 1. Molecular Canvas

You can now **draw molecular structures directly or enter SMILES to create and edit molecules in chat, then use them together with prompts for research requests**.

#### Add Molecules by Drawing Structures

* Open the molecular canvas from **\[+] → Draw Molecule** in the prompt input box.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2F6FmJd0LawpcxBCCg7DXY%2Fimage.png?alt=media&amp;token=f707c8ae-f8dc-45a4-a3ee-ded3251fddb3" alt=""><figcaption></figcaption></figure>
* Draw a molecule directly or enter a SMILES string to create a new structure.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2FO7DLwUGd5eWjGQq5RJog%2Fimage.png?alt=media&amp;token=43a230fa-d35c-43c3-af48-dab63be46d6f" alt=""><figcaption></figcaption></figure>
* Add multiple molecules to the prompt input box and compare or analyze them in a single request.
* Reopen molecules added to the input box to edit their names or structures.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2FSxF8bPZ3kvrtJylRFWTF%2Fimage.png?alt=media&amp;token=21137dc8-65fe-4bf2-bdca-35495d0670ca" alt=""><figcaption></figcaption></figure>

#### Edit Molecular Structures Shown in Chat and the Viewer

* Modify SMILES strings or 2D molecular structures shown in previous chats using Molecular canvas and use the edited structures in new research requests.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2FRY1u1IUzWaBKoZ87vxra%2Fimage.png?alt=media&amp;token=9f168982-ae19-4140-8686-143bd39ec76f" alt=""><figcaption></figcaption></figure>
* In the Library molecular structure Viewer, you can also edit structures using **\[Draw Molecule]**.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2FmYWHNtlc0tSmSOTiuuAg%2Fimage.png?alt=media&amp;token=ef4747f6-daec-4f20-8a04-02fb3e7dde9d" alt=""><figcaption></figcaption></figure>

For example, you can modify a specific functional group or atom in an existing candidate compound and request a comparison of **changes in activity or molecular properties resulting from the structural modification**.

Using the molecule drawing feature does not directly modify the original chat results or source files in the Library. Edited molecules are added to the prompt input box as new inputs for subsequent research requests.

### 2. Protein–Ligand Binding Affinity Prediction

A new capability has been added to **predict binding affinity between proteins and ligands**.

Depending on whether the researcher already has a **protein–ligand complex structure**, binding affinity can be predicted in two ways.

<figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2F21tN8soKbOcozajJdeDE%2Fimage.png?alt=media&amp;token=18489e63-ed19-468d-bea9-1163825e24e5" alt=""><figcaption></figcaption></figure>

* **When the complex structure is not available**\
  K-Fold can be used to model the 3D structure of the protein–ligand complex and estimate binding affinity based on the predicted structure.
* **When the complex structure is available**\
  An existing protein–ligand complex structure can be provided as input to predict binding affinity using **PIGNet**. PIGNet is a **deep learning model that incorporates physics-based information** and predicts binding affinity by considering interactions between the protein and ligand.

This allows researchers to evaluate binding affinity using an approach that fits their research stage, from early-stage studies where no complex structure is available to later stages where an experimentally or computationally obtained structure is already available.

You can run the analysis by asking Bio Co-Scientist to calculate binding affinity, making it easier to compare and evaluate the protein-binding potential of candidate compounds throughout the research process.

### 3. HWP Document Viewing in the Library

You can now **view HWP and HWPX documents uploaded to the Library directly in the Library Viewer without using a separate application**.

<figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2F7LU3VtuQYHwt30bFjOMg%2Fimage.png?alt=media&amp;token=45db6d77-0763-47f2-8af8-4731392f0f64" alt=""><figcaption></figcaption></figure>

* Both HWP and HWPX formats are supported.
* Documents can be viewed in the Viewer, but editing is not supported.

This makes it possible to open and review Korean research documents directly without downloading or converting them first.

### 4. Collapse and Close Controls for Question Cards

The usability of question cards used by the Agent to request additional information during research has been improved.

* Use **\[Collapse]** to minimize an open question card and review previous chat content.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2FVChI8krTz9le8MUqT6QT%2Fimage.png?alt=media&amp;token=0e446c9c-c2bf-4418-9016-5a6ff4879e8c" alt=""><figcaption></figcaption></figure>
* If you want to continue with another request without answering the current question, you can **\[Close]** the question card.
* Closing the question card reactivates the prompt input box so you can immediately enter a new request.

  <figure><img src="https://709489233-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUd87mch9xATD0InicbJH%2Fuploads%2Fj9RAvmmgrb6yNRiB3D1y%2Fimage.png?alt=media&amp;token=99bb3aa3-b2c8-449a-adbc-15813e1bd947" alt=""><figcaption></figcaption></figure>

This makes it easier to review previous research before answering a question or skip the current question and **switch to a different research request**.

***

## ⚒️ Bug fix

Bug fixes and overall service stability improvements have been applied.
