Hippocampome.org Synapse Knowledge Base

Requirements:

Easy access:

(1)              To access data, please open the following link and add the "synapse" folder to your own Google Drive. You need to do this only once.

(2)              Go to the "synapse" folder and make a copy of the "Hippocampome.org/Synapse" and "FormTemplateReview" files. You need to do this step only once as well. The copied files will appear in the root of your drive, so you need to go back to "My Drive" to access these copies.

When you open the "Copy of Hippocampome.org/Synapse" spreadsheet, you should see a new "Data Mining" menu that we have custom-built.

The Google Sheet you just copied has different tabs, some of which might be hidden. You can access them by going to "View" -> "Hidden sheets."

The list of tabs:

How to access the search engine from Google Sheets

To get the latest updates on the synapse type mapping,

(1)              Select the "Evidence" tab.

(2)              Select an experiment, i.e. a block of non-empty rows. Rows 11 to 17 that are selected in the following figure, for example, belong to one experiment. The first four columns show the synapse types mapped to the experiment.

(3)              Select "Check Query" from the Data Mining menu as shown in the following figure.

(1)(3)To use the Data Mining tool, grant execution authorization and press "Continue". You need to do this only once.

Use your account that has access to the "Synapse" folder and the Google Sheets copy.

On the next menu, press "Advanced" as highlighted in the following figure.

Press "Go to Automation." "Automation" is the source-code package name backing the Data Mining tool.

Press "Allow."

When the confirmation menu appears, press "OK."

The search result should open automatically. The tool compares the current search engine results with mapping results at the time of data mining in order to suggest new synapse types to be added (table with blue header) or deleted (table with red header). You can also refine the search query and repeat the search, save the new query, and sort the tables by clicking on the column headers.

How to review an experiment and linked references:

(1)              Select the "Evidence" tab of the sheet (see figure below).

(2)              Select an experiment.

(3)              Select "Review References" from the Data Mining menu.

Wait a few minutes for the review panel to pop up (refer to figure below).

References are editable but saving them is not automatic. You need to press save button (💾) yourself. After a change the save button will start blinking to remind you that you need to press save. The red ❌ button removes the HTML text formatting. The blue icon next to it opens a menu that assists typing Unicode characters. The following shortcuts are supported:

Text excerpts are highlighted automatically in different colors to assist data extraction and referencing.

To review or modify the patterns, find the "refHighlights" function in the "Script Editor" (see advanced access).

Advanced access

To access forms,

(1)              Select "Tools" -> "Script editor" from Google Sheets.

(2)              Find the "DataMining.gs" file that automatically opens.

(3)              Set the proper ID values for the "reviewFormID," "dataExtractionFormID," and "synapseSpreadsheetID" variables. You can find these IDs in the URL of the "Copy of FormTemplateReview," "Copy of FormTemplateSynapticData," and "Copy of Hippocampome.org/Synapse" files, respectively.

(4)              You can also change the values of "EZProxyLink" and "EZProxyName," with those representing your own institute.

(5)              Select the "setup" function and press the "run" button. You need to do these steps only once.

Now you can go to the Google Sheet, select an experiment, and from the "Data Mining" menu either select the "Review Evidence" or "Extract Data" buttons.

Confirm your selection.

Wait a few minutes for the panel to open with the "review form" and references panels side by side. You can do any edits you want to the form and then submit it. All changes will automatically be saved in the appropriate rows in the Evidence or Covariates tabs, on pressing the submit button. Be careful not to add or remove rows before submitting the form. If you are compelled to do so, you can appropriately edit the "Active Range" value.

If you select "Extract Data," the data extraction menu opens. The confirmation menu is different here. You can choose to preload the data extraction form, with data extracted already. You need to know the data ID (dID) of the extracted data row for this purpose, for instance, "2" in the following example. You can also choose to not enter the dID in order to start from scratch. In this case, the program will do its best to prefill the form using information already available in the Evidence and Covariates tabs.

Press OK and wait a minute to see the prefilled form.

On pressing the submit button, the extracted data will be saved in the SD tab.

Reversal Potential and Liquid Junction Potential Calculator

To test the calculator:

(1)              From the Google Sheet go to "Tools" -> "Script editor."

(2)              Open the "Erev.gs" file.

(3)              Uncomment one of the solutions.

(4)              Press the "run" button.

(5)              Go to "View" -> "Logs" to see the results.

Notes:

Text cleaner tool

Assume you want to make a new reference for a text excerpt from a PDF file.

Copying and pasting the text will yield a lot of unnecessary "end of the line" (EOL) and "ASCII control" characters. Paragraph marks in the below figure show some of these hidden characters.

We need to delete all of the unnecessary EOLs and correct misspellings like "0 2," which should be "O2". The "Text Cleaner" tool largely automates this process. We have discovered the most common error patterns and detect them with proper regular expressions. The text cleaner tool not only helps text extraction from PDF sources, but also from HTML. Some HTML pages use "small-caps" characters that appear capitalized on the web pages, but, when pasted, turn to lower case. For instance, the "mM" text that is copied from the Journal of Neuroscience will convert to "mm" on pasting. Our tool reverts this unwelcome conversion. Many of the remaining spelling errors are recognized by the browser’s built-in spellcheck functionality. To access the text cleaner,

(1)              Go to one of the tabs assigned for the excerpt storage, "Da" for example. Click on an empty cell under the "Excerpt" column,  then press "Data Mining" -> "Text Cleaner."

(2)              Copy and paste the text you want to clean in the tool window and press save. Common problems in the text will be corrected, all unnecessary EOLs will be deleted, and the clean text gets saved in the selected empty Google Sheet cell.

(3)              To link the text cleaner tool to another spreadsheet cell you can select another cell and press the 🔄 button.

Text Cleaner also gives you access to a list of Unicode characters often used in scientific reports.

In your own data mining endeavors, if you find a new recurring pattern that you wish to be automatically corrected, or if you want to modify, disable, or enable the existing patterns, you need to edit the "TextCleaner2.html" file. Each "replace" command in the "formatText" function controls one of the patterns.

Data format

We have designed a human- and machine-readable format to store data entities inspired by common author usage. A single data entity in string form will look like this:

V1 ± V2 [V3 to V4] (n=V5) {inner comment}@V6 {outer comment}

where,

Multiple data entities can be linked to a RefID set. In this case data entities are joined with semicolon. Data entities with different excerpts are joined with comma, as in the following examples.

1.4±0.74(n=7){[Ca²⁺]ₒ=1}; 3.85±0.71(n=7) {[Ca²⁺]ₒ=2}@3200835{digitized and calculated}

-65.2±17.2(n=7){septal}; -65.0±8.3(n=17){temporal}@3201035{PVBC→dPC:High [Cl⁻]ᵢ}

-15.2±1.8 [-26.7 to -12.8](n=4)@3201032{CCK→PV}, -12.8±0.4(n=1)@3201033{CCK→PV}

-32{without Vj correction}; -49{with Vj correction}@3200666&3200667{Vh}

Data parser

Our custom JavaScript parser converts data to JSON or "stringify" them again. To access the parser, download the "parser.js" file in the synapse folder. In R, for instance, we use Google’s V8 JavaScript engine and our parser to convert the data.

library(dplyr)
js <<- V8::v8()
js$source("parser.js")
js$call("dataParser", "1.4±0.74(n=7){[Ca²⁺]ₒ=1}; 3.85±0.71(n=7) {[Ca²⁺]ₒ=2}@3200835{digitized and calculated}") %>%
    jsonlite::toJSON() %>%
    jsonlite::prettify()

The previous examples in JSON format:

[

    {

        "RefIDs": [

            "3200835"

        ],

        "note": "digitized and calculated",

        "values": [

            {

                "v": 1.4,

                "s": 0.74,

                "n": 7,

                "note": "[Ca²⁺]ₒ=1"

            },

            {

                "v": 3.85,

                "s": 0.71,

                "n": 7,

                "note": "[Ca²⁺]ₒ=2"

            }

        ]

    }

]

 
[
    {
        "RefIDs": [
            "3201035"
        ],
        "note": "PVBC→dPC:High [Cl⁻]ᵢ",
        "values": [
            {
                "v": -65.2,
                "s": 17.2,
                "n": 7,
                "note": "septal"
            },
            {
                "v": -65,
                "s": 8.3,
                "n": 17,
                "note": "temporal"
            }
        ]
    }
]
 
[
    {
        "RefIDs": [
            "3201032"
        ],
        "note": "CCK→PV",
        "values": [
            {
                "v": -15.2,
                "s": 1.8,
                "ll": -26.7,
                "ul": -12.8,
                "n": 4,
                "note": ""
            }
        ]
    },
    {
        "RefIDs": [
            "3201033"
        ],
        "note": "CCK→PV}",
        "values": [
            {
                "v": -12.8,
                "s": 0.4,
                "n": 1,
                "note": ""
            }
        ]
    }
]
 
[
    {
        "RefIDs": [
            "3200666",
            "3200667"
        ],
        "note": "Vh",
        "values": [
            {
                "v": -32,
                "note": "without Vj correction"
            },
            {
                "v": -49,
                "note": "with Vj correction"
            }
        ]
    }
]

 

We use the parser and a set of customized helper functions to analyze the data and calculate conductance potencies in R. The file "SynapticDataAnalysis.Rmd" in synapse folder gives access to the R source codes.

How to give us attribution

If you found our data or these tools useful in your own research, please cite our work: Moradi and Ascoli, 2019 A comprehensive knowledge base of synaptic electrophysiology in the rodent hippocampal formation. Hippocampus 2019 (in press); doi: https://doi.org/10.1101/632760.