Hi! From your description, it sounds like you're doing everything right in terms of checking for loading consistency, using Ponceau staining, validating protein quantification with BCA, and comparing multiple loading controls (β-actin and GAPDH). However, it’s not uncommon to encounter inconsistencies in band intensity that don’t align perfectly with total protein loading, even when everything appears to be technically sound.

A few points might help explain what you're seeing:

Suggestions:

• Consider using a total protein stain (like REVERT or Stain-Free) as a more quantitative and consistent loading control — these often outperform single-protein markers in terms of reproducibility.
• If variability persists, normalize your bands to the Ponceau signal using densitometry; it’s not perfect, but better than relying solely on a fluctuating housekeeping protein.
• Try running all replicates on the same membrane when possible to reduce between-blot variability.

Hope this helps! It sounds like you’re already being very thorough, which is great, western blots can be finicky, and these kinds of inconsistencies are frustratingly common. Good luck, and let us know if you discover anything else in troubleshooting!