This "gap" are simply some extraneous pixels from the edges. Pre-compose to crop accordingly.
pre-composing doesn't seem to do anything
It is common for video to have a column that is two or three pixels wide on the left and right side of the frame that are empty (zero value) or black. These are hard to see and are very common with older video cameras and quite common with film transferred to tape. This will put a black wedge at the top of the circle when Polar Coordinates are applied to the footage and the distortion is set to 100% and the conversion is Rectangular to Polar.
Here's what i think is going on in your screenshot. I think you have the Motion Tile FX turned off in the time line, because if it was on then the black wedges would be extending left and right from the center and be vertical. Your transform effect is redundant because you could scale the footage to 100, -100 with the Transform>Scale and achieve the same thing.
To check for the typical black pixels at the left and right edge of video simply turn off the Polar Coordinates effect and check just motion tile with width set to 50%. You should see two vertical gaps in the image. Come to think of it, if you were to adjust the Tile Center to half of the movie width or twice the movie width then you could move the wedge to the vertical. If you do not see two vertical black lines or empty pixels then let us know.
The solution to your problem is to scale up your movie layer to eliminate the empty columns of pixels left and right, then pre-compose moving all attributes, then scale the pre-comp 100,-100, then apply Motion Tile with width set to 50, then apply Polar Coordinates.
A more informative screenshot would have included the project panel with the footage selected so we could see the properties of the source file and you would have selected the footage layer and pressed the U key twice to reveal all changed properties in the layer. The whole story of your comp would be much easier to understand then. BTW, if you're using 8 bit effects like motion tile, you should also be using the commander to keep the high bit rate advantage in your project. This is how I would do it. I used a shape layer that was 10 pixels narrower than the comp to simulate the empty pixels in the video. The layer was scaled up in the pre-comp, then the HDR Compander to Compress, Motion Tile, the HDR Compander to expand, then Polar Coordinates were applied.