Nicole MRes Project README

All code used for EthoVision behavioural analysis and Keypoint-MoSeq behavioural analysis pipeline is available on the lab GitHub account.
Larger files, such as the RNAscope images and videos for Keypoint-MoSeq , can be found on the University long-term storage, either by clicking the link below or via Map Network (X:\nj35\MRes). For RNAscope, all images captured using Zeiss 3.9 software on Jeeves PCs, stored in Documents/Nicole/ and are uploaded to the long-term storage.

Allodi Lab GitHub Repository:
https://github.com/Allodi-Lab

University Long-Term Storage:
file://cfs.st-andrews.ac.uk/shared/AllodiLabs/ (for Windows)

---

Language

• Behavior_Analysis – Keypoint-MoSeq (Python + R) – EthoVision (R)
• RNAscope_Analysis – (GraphPad Prism)

---

Version

• DeepLabCut + Model Zoo Package – 3.0
• Keypoint MotionSequence – 0.5.0
• GraphPad Prism – 10.7

---

Note:
For Ethovision_analysis, Ethovision outputs (raw) were compiled into excel files for each group (control, post_w1, post_w2) for eac paradigm. The summary excel files were used for analysis, which is a compilation of the raw excel files. model statistics were first determined, then raincloud plots were produced, which were used in the thesis.
For the MoSeq_analysis, model statistics were first determined, then the SE values were determined. The SE values were then ordered by control frequency. The ordered_SE data was used for the thesis. GEELM model was just for testing and can be ignored.
For RNAscope_analysis, data used in the thesis was analysed in GraphPad Prism, using a Nested T-Test. blind excel file is the raw values from Zeiss, renamed. macros excel file is the averaged excel file, aided by excel macros for automation of calculations. processed excel file formats the values easier for GraphPad Prism, with trnapose excel file being the same file but transposed. Example images, which were also reported in the thesis, can be found under the Zeiss folder. The cropped versions of the behavioural videos were used for DeepLabCut + Keypoint-MoSeq to prevent model training on paradigm reflections. There are only folders for Cohort 1 + 3, as cohort 2 only had one mouse and so was combined into cohort 3. Download .html files to view all outputs for R-coding notebooks.