The configuration system in DeepVocal has strict limitations. You can only have -CV, CV, V_V, V_C, and V_tail. Independent symbols no longer function properly in the current toolbox.
Any symbol categorized as vowel must have a pitch (ie. No whispering or growling or noisy vocals)
Any symbol categorized as a voiced consonant must also have a pitch (ie. Consonants like m and n which you can sing notes with. Consonants like z are often too noisy for DeepVocal to accept)
With those limitations in mind, you can develop an equivalent to a CVVC voicebank in UTAU. You could also skip the other configs and only include CV, if you want to ignore the warning messages while building the voicebank.