
JelloX Biotech Inc.
# MetaLite User’s Guide
## **Quick start**
:::spoiler If you want to quick know how to operate MetaLite, you can unfold this section and follow the instruction below.
### Open file


### Change view
|Zoom in/out |Drag move|
|-----------|---|
|||
### Edit annotation

Use ```Ctrl``` + ```Mouse Left Buttom``` to adjust drawing size, and press ```Mouse Left Buttom``` to edit annotation.

### AI assisted annotate
User AI model to inference the tumor area.

### Antibody statistic
Statistic the antibody expression of annotated area.

:::
## Introduction
MetaLite software developed by JelloX Biotech Inc. is a Research Use Only (RUO) computer-assisted image analysis software. It is intended to aid pathologists and researchers in imaging annotation and analyzing cells of interest based on marker intensity, size, shape and volume using appropriate tools. This software also assists the users to select fields of view (FOVs) in the digital image with an embedded artificial intelligent based feature recognition model, and provides quantitative data on these FOVs such as marker distribution, distance between markers, and ratio of different markers.
<div style="text-align:center" markdown="1">

</div>
### Supporting file format
Types of Files That You Can View.
MetaLite can load and view different types of data, including:
- JPEG and JPG files
- TIF files
- Multi-Layer TIF files
- OME-TIFF files [(Open Microscopy Environment)](https://docs.openmicroscopy.org/ome-model/5.6.3/ome-tiff/)
- PNG files
- SVS files
- NDPI files
- DICOM files
## Installing
This chapter contains information of installing the MetaLite software
### Before You Start
Review the information in this section prior to installing MetaLite.
### Requirements
OS
- Windows10
CPU
- Intel Core i3 6300 or Equivalent
GPU
- Lower then NVIDIA GeForce® RTX 20 Series
### Installation
To install MetaLite, follow these steps:
1. Ensure you are logged into Windows as a user with administrative privileges.
2. Double-click “My Computer” or open “Windows Explorer” and navigate to the MetaLite installer file. (Please [subscribe MetaLite](https://jarvislin1.wixsite.com/metalitesubscribe) or [contact us](https://jellox.com/#contact).)
3. Double-click the “.msi” file to start the installation wizard.
4. Follow the instructions on your screen to accept the terms of license agreement to complete installation.
### Modifying or Removing the MetaLite Software
Once MetaLite is installed, you can run the installer again to modify, repair, or remove the MetaLite software. If MetaLite is already installed, select from the following options on the installer window:
- **Modify** to change the MetaLite installation by adding or deleting components.
- **Repair** to reinstall all the components previously installed. This is the option to use if you are upgrading a previous installation to new software.
- **Remove** to uninstall the MetaLite software.
### Starting MetaLite
To start MetaLite, click **Start** on the Windows taskbar, point to **All Programs** > **MetaLite**, and select **MetaLite**. Or click the **MetaLite Icon** on the Desktop.
## Overview
This chapter provides a tour of the MetaLite main window and describes how to use the navigation and magnification tools.
### Main elements:
- **Main viewer** – The main viewer of the image.
- **Menu** – The menu bar lists all functions of MetaLite by categorizing them into File, Edit, Image, View, Tools, Windows, and Help. Some of the functions are locked in this version, please contact us to un-lock more details and advanced functions.
- **Toolbar** – The toolbar contains a list of annotation and statistic functions. See next section for a quick reference list of each MetaLite toolbar icon.
- **Zoom slider** – The zoom slider can magnify or shrink the current view.
- **Image slice controller** – The image slice controller appears with multiple images only, and is used to view different focal areas on a z-stack image.
- **Thumbnail window** – The thumbnail is a navigation tool that shows complete Image.
<div style="text-align:center" markdown="1">

</div>
#### Main viewer
The main viewer displays the image with a region of interest (ROI). To move the image, press right mouse button on the Main viewer, and drag the image to another ROI while the cursor turns into grasp hand.
<div style="text-align:center" markdown="1">

</div>
### Channel staining
<div style="text-align:center" markdown="1">

</div>
- **Staining type** – The staining style of main viewer shows, include:
- RGB
- Pseudo HE
- Pseudo IHC
- Pseudo staining
### Statistic panel
- **Data value grid** – The form of statistic data.
- **Data chart** – Plot the value in the data value grid.
<div style="text-align:center" markdown="1">

</div>
### Menu
#### Open File
- **Open File** ```Ctrl + O``` – images that are stored on your computer.
- **Open Annotation** – Open the ".jellox" file that pre-annotated with Labels of the image. The annotations labeled by MetaLite will be saved as a ".jellox" file. When loading the image, if there is a same name ".jellox" file in the same file directory, MetaLite will automatic load the annotation file. And you can also read ".png" as annotation, it turns the non-zero pixel of png alpha channel to annotation.
<div style="text-align:center" markdown="1">

</div>
#### Save Annotations
- **Save Annotation** ```Ctrl + S``` – Save the labeled annotation layer of current image to the ".jellox" file.
- **Save Annotation As** ```Ctrl + Shift + S``` – Save the annotation to a file named alter to current Image name.
- **Save All Annotation** ```Ctrl + Alt + Shift + S``` – Save all the annotation layers of each image at once.
- **Export Current Annotation** – Export the current selected annotation as ".png" file.
- **Auto save** – Switch on/off the auto Save Annotation method to period execute the Save Annotation to preserve work stage.
<div style="text-align:center" markdown="1">

</div>
#### Activate analyzer
- **Activate MetaLite Analyzer** – The statistic calculation of the MetaLite software is working into a virtual machine called "Metalite Analyzer". If the server is disconnected or unavailable, you can restart the MetaLite server manually. Once the analyzer is activate the JelloX logo will turns from black white into colorful.
<div style="text-align:center" markdown="1">


</div>
## Tools
### Annotation tools
| Tool | Action |Shortcut|
| :--: | ----------- |:---:|
||**Draw**: A free-form annotation pen drawing tool that labeling the annotation area. |D|
||**Eraser**: A free-form annotation erasing tool that can eliminate the annotation area we don’t want. |E|
||**Polygon**: A polygon annotation region drawing tool.|P|
||**Snapshot** A snapshot tool that can save selected retangle region as ```.tif``` image file.||
:::info
**Tips:**
- **Draw and Erase** – Use **Ctrl + MouseWheel** to adjust the drawing/erasing brush size. Use Shift + MouseWheel to quickly adjust the transparency of annotating brush.
- **Polygon** – Press **Esc** to cancel selection, and press **Enter** to close the selected contour. Double click the main viewer to close the selected contour, or click to the start point.
:::
### Statistic tools
| Tool | Action |
| :--: | ----------- |
||**AI Analysis** can automatically annotate the cancer tissue area. <br/>We provided Fluorescence image and histopathology image inference model, and the inference process are accelerated by Cuda and Openvino hardware framework. |
||**Cell Counting** can counting objects of specific channel on fluorescent image.|
||**Expression** of antibody signal percentage that is covered by current annotation layer. The demo statistic method in this version is for nuclear-specific antibody percentage calculation. (If you are interested in other statistic method such as membrane-specific antibody expression, TIL calculation, please contact us for more information.) |
||**Export Report** function can help pathologist product pathology diagnosis report, and can attach the executed **Expression** value below the report. |
## Main operation
### Main viewer zooms in/out.
Scroll the Mouse Wheel up/down on the main viewer to zoom in/zoom out the image view.
<div style="text-align:center" markdown="1">

</div>
### Scale sliding bar of the image view
Drag the sliding bar next to the thumbnail view or there is a quick scale button next to the sliding bar.
<div style="text-align:center" markdown="1">

</div>
### Annotation layer panel
- **Add Annotation** – Use to add new annotation layer to current image.
- **Annotation Name** – Double click to rename the **Annotation Name**. The gray bar in the background represents the transparency of that annotation layer.
- **Color Selector** – Select the annotation color.
R – Red intensity
G – Green intensity
B – Blue intensity
A – Alpha intensity, the inner annotation transparency.
<div style="text-align:center" markdown="1">

</div>
- **Delete annotation** – Delete selected annotation layer.
- **Annotation Switch** – Show/Hide the annotation layer. Notice, even the annotation is hidden, the Draw and Erase functions can still work on the annotation layer.
- **Hide annotation board** – Hide/Show the annotation board and thumbnail.
<div style="text-align:center" markdown="1">

</div>
## Function Detail
### Hide all annotation
Hide all annotation layers to quickly check the tissue covered by the annotation.
<div style="text-align:center" markdown="1">

</div>
### Image staining
Change the staining algorithm of each channel of the image to RGB/Pseudo IHC/Pseudo H&E/Pseudo Staining color-code.
<div style="text-align:center" markdown="1">

</div>
You can drag the stain board to adjust the displayed color on the screen.
<div style="text-align:center" markdown="1">

</div>
In pseudo H&E, you can use left button to drag staining board to change the channel staining. And modify the staining color to other color.
<div style="text-align:center" markdown="1">

</div>
Double click on the slider thumb to choose color.
<div style="text-align:center" markdown="1">

</div>
## AI analysis
We use pre-train AI model to predict tumor region. Once you press the AI analysis button, it will create a new annotation layer and label the tumor region on the layer.
:::info
**Constrain:** For some higher version compatibility of cuda and hardware affinity, the GPU higher then RTX 30 Series are unavailable to execute AI model inference.
:::
### Fluorescence model
This model was trained on our fluorescence breast tumor image, and staining nuclear with red channel, membrane with green channel.
<div style="text-align:center" markdown="1">

</div>
### Histopathology model
This model was trained on camelyon breast lymph node image.
<div style="text-align:center" markdown="1">

</div>
:::info
**Notice:** This version for inference 5000x5000 image on AMD Ryzen 5 3600X Six-Core Processor computer will take about 30~40 seconds. It could be accelerated by Nvidia GeForce RTX 2080Ti 11GB GPU to 3~4 seconds. And the Histopathology image model are running on openvino framework, it's only supported intel CUP hardware.
:::
## Image tool
### Format convert
For some of the specific WSI format we provide a function that can convert the image from hardware/software affinity format to common open source format.
1. **Image>Format Convert**


2. Select the input image.

3. Select output file.

4. Click convert button.

### Supported format
#### Input
- Philips iSyntax
#### Output
- OME-TIFF
## Statistic
### Cell Counting
We can count the object amount of specific channel of the image.
1. Select the channel that represented the object we aim to counting. 
2. Select the image apparent magnification 20x/40x, set the cell pixel size.
3. Click **"OK"**.

### Expression
We calculate nuclear-specific antibody covered percentage area of the nuclear counter stain to measure the expression level (High/Low then 14%).
1. Select the color channel corresponding to each cell component.

2. Select the image apparent magnification 20x/40x, set the cell pixel size, and threshold of cell channel intensity to optimize the algorithm.
3. Click **"OK"**.

### 3D View
If your data are sequence images of tissue, it can be represent as 3D object by our 3D viewer function.
1. Make sure multipage datas are loaded.
2. Click the 3D viewer function.


## Contant us
- Official website:
- https://www.jellox.com/
- Phone:
- +886-3-5620829
- E-Mail:
- jarvislin@jellox.com
- changsy@jellox.com
- Address:
- No. 66-5, Shengyi 5th Rd., Zhubei City, Hsinchu County 302041 , Taiwan (R.O.C.)
- 302041 新竹縣竹北市生醫五路66之5號