Multiple drugs were applied to mutant fish. Identify the one drug that will make the fish behave the most like the wild type.
- calculate zscore based on wild type, calculate_zscore.m/calculate_zscore_burst.m
- average across all the fish for each drug and geno (HOM and WT), average_after_zscore.m
- make clustergram, avgz_to_clustergram.m
- PCA and euclidean distance.
More Details: The following two process lines work:
Individualdata ---> calculate_zscore ---> average_after_zscore ---> avgz_to_clustergram
splitmean ---> calculate_zscore_burst ---> average_after_zscore ---> avgz_to_clustergram
This one needs more work:
splitmean ---> calculate_zscore ---> average_after_zscore ---> avgz_to_clustergram
| File name | Description |
|---|---|
| calculate_zscore.m | Input type 1: Jeff's pipeline (usualy named 'individualData_xxxxxxxx.csv') Output: calculate the z-score of experimental fish plate based on the mean and standard deviation of the wild type fish of the same day Input type 2: The one off burst file generated by Jeff (named 'scn1lab_rw_split_means.csv') Output: calculate the z-score of experimental fish plate based on the mean and standard deviation of the wild type fish of the same day, TO DO: removed burct and burur coloumns because there are too many infs |
| calculate_zscore_burst.m | Input: The one off burst file generated by Jeff (named 'scn1lab_rw_split_means.csv') Output: calculate zscore of the burst variable (burct and burur), but use wild type from all days because each single given day the fish may not have any burst activity. |
| average_after_zscore.m | Input: calculated z score from the above 'calculate_zscore.m' and 'calculate_z_score_burst.m' Trim the file so only HOM types are left Average across fish, also aggregate across activities Output1: mean_by_geno. The parameter/activities are averaged acrossed fish for each geno If it is not the bust file, also export the aggregated zscore: Output2: _averaged. The parameter/activities are aggregated to rms and mean for each group, bout, activity, sleep, all |
| avgz_to_clustergram.m | Use the mean_by_geno to generate a clustergram. Use customer color my_colormap. Replaced the underscores in the labels with space. |
| run_PCA_euclidean.m/td> | Input, mean by geno file. Output: graphies for publication, euclidean distance table. Run PCA, make plots for publications, and calculate euclidean distances for all the drugs and dosages |
| script_best_recovery_drug.m | Combining the above steps to plots and calculate which drug is the best recovery drug. |
Fish in 8 rows (96 plates total)
PRE POST
Wild Type — DMSO Wild Type — DMSO — H2O
Wild Type — DMSO Wild Type — DMSO — PTZ
Wild Type — MC Wild Type — MC — H2O
Wild Type — MC Wild Type — MC — PTZ
HOM — DMSO HOM — DMSO — H2O
HOM — DMSO HOM — DMSO — PTZ
HOM — MC HOM — MC — H2O
HOM — MC HOM — MC — PTZ
- PTZ creates more seizure (burst count)
- effect of PTZ in HOM and Wild Type(WT)
- effect of MC (Can MC cancel PTZ's effect)
- Take the pre and post difference score, and do a 3-way ANOVA, geno x MC/DMSO x PTZ/H2O, look for significant 3-way interaction.
- Separate HOM and WT, and do 3-way ANOVAs: pre/post x MC/DMSO x PTZ/H20 Look for significant 3-way interaction only in HOM but not in WT
- post hoc t tests: compare HOM+MC+PTZ and HOM+DMSO+PTZ, they have a similar value pre, but in the post data, HOM+DMSO+PTZ has a higher value.
From Raw score for pre and post experiments to various plots, anovas, and an intermdeidate csv file
- Step 1: first make a folder that includes the following three files: pre, post, geno.
- Step 2: save the pre and post files into the excel format, with the xlsx extension. (Matlab is able to easily convert excel files into tables, but not from csv files)
- Step 3: run 'script_pre_post_analysis;' on the command line. You will select the three files in order: pre (the excel version), post (the excel version), geno. If it's an windows computer it would show the prompt for each type of file, but the prompts don't work in a Mac OS system. so just follow the order of pre-post-geno.
- Step 4: check the outputs which will be saved in the folder that you have created in the first place!
| File name | Description |
|---|---|
| pre_post.m | preprocess the pre and post data files and geno text file, generate an output that can feed into the following analyses |
| plot_geno_by_time.m | use the output from the pre_post and make geno_by_time plots, saved in a 'plots' folder |
| plot_prepost.m | plot the pre post data |
| plot_bars.m | plot the data by all the combinations of factors |
| make_boxplots.m | make boxplots for all the combinations of factors including geno, drug1, drug2, pre and post |
| do_anova.m | do 3-way anova on the difference score (geno, drug1, drug2), HOM and WT data (pre/post, drug1, drug2) |
| script_pre_post_analysis.m | a script that chains oher functions. First use pre_post to create the output variable. Then use the output variable to make various plots and do anovas. |