Use the same method that used in PepQueryMHC (https://github.com/bzhanglab/PepQueryMHC).
- Introduce class level scoring.
- Worst case as a representative category per transcript.
- If there are multiple transcripts in the same region, select best transcript based on penalty rule.
Use the same method that used in PepQueryMHC (https://github.com/bzhanglab/PepQueryMHC).