The reference to Pub 590 (which clearly needs to be updated by Lacerte, it's been several years now since the IRS split 590 into A and B) is likely a reference to Pub 590-B Worksheet 1-1 (page 17 of the 2020 Pub).

https://www.irs.gov/pub/irs-pdf/p590b.pdf

It's referenced on page 15 under "Contribution and distribution in the same year."  When I've run into these there's been a lot of jumping back and forth between the 1040, the Pub and the 8606.  I know ProSeries data entry is quirky on this and you have to make sure you enter everything in the right boxes for the math to come out correctly.  I seem to recall for at least one of the worksheets there was an important distinction between a 2020 IRA contribution made before 12/31/20 (therefore included in the 12/31/20 FMV) and one made for 2020 in 2021 by 4/15 (not included in the 12/31/20 FMV).  I would guess Lacerte is also particular about what needs to be entered where to get everything in the right place on the forms/worksheets.

Any time I don't like what I'm getting out of the software I go back to paper forms and worksheets and work through the calculations by hand.  Then you should be able to identify a particular box/line that derailed you and try to figure out what the software wants you to do.

Beyond that, great info from Bill (as usual).  I don't know of anything that would subtract Roth conversion income from AGI to determine MAGI for IRA contribution deduction purposes.

As you work through the worksheets, pay close attention to where the basis goes for the non-deductible contribution.  Does it stay with the "new" IRA or does part of it get allocated to the Roth Conversion income?  (Effectively moving your IRA deduction from Schedule 1 Line 19 to the 1040 Line 4b, which may have little effect on your Taxable Income.)  Not enough data for me to dig into the details but this may be a case where the IRS is adjusting S1-L19 and you need to convince them to make an adjustment in the opposite direction to 1040-L4b.

Rick