search

MassBank / MassBank-data

Official repository of open data MassBank records
74 stars 59 forks source link

Dev: RIKEN CSRS records #74

Closed zzjl20 closed 5 years ago

zzjl20 commented 5 years ago

10239 records form RIKEN Center for Sustainable Resource Science

meier-rene commented 5 years ago

Hi, I had a look in your contribution and could fix some issues myself. One of the points is the usage of MS$FOCUSED_ION: ION_TYPE. Recently we had a discussion about the usage of this tag in https://github.com/MassBank/MassBank-web/issues/176 and came to the agreement, that it is better to use MS$FOCUSED_ION: PRECURSOR_TYPE for MS2 spectra. I changed this for all your records and I would appreciate it if you could change your code accordingly. Another point is the CH$FORMULA tag. If the measured molecule is already an ion than we expect something like [C42H69O12]-. Any chance to introduce this in your future contributions?

But there is another major thing I can not figure out: some spectra are positive mode and some are negative mode. But the adduct type is not in agreement with the ion mode.

I found 797 records with ION_MODE NEGATIVE and a positive adduct ion. example PR311184.txt I found 4216 records with ION_MODE POSITIVE and a negative adduct ion. example PR308655.txt

Do you have any explanation for this?

zzjl20 commented 5 years ago

Thank you for the help of MS$FOCUSED_ION: PRECURSOR_TYPE for MS2 fix. I made a note and will pay attention for my work. For the CH$FORMULA, sorry I didn't notice the format changed on April 25th. I think I can solve this. Ion_mode maybe is something wrong. I need to find out.

meier-rene commented 5 years ago

Thank you for your working on the data, it looks better now. Nevertheless the problem with the mismatch of ion mode and precursor type is not solved. I can see that you changed a lot of files but sometimes you changed it from 'correct' to 'incorrect' like in PR300001.txt. It was ION_MODE POSITIVE and PRECURSOR_TYPE [M+H]+ and now its ION_MODE NEGATIVE and PRECURSOR_TYPE [M+H]+.

I will attach a list with mismatches between ION_MODE and PRECURSOR_TYPE: ionmode.txt

There are over 8000 records with this problem. Please correct them.

zzjl20 commented 5 years ago

Feel frustrated... OK, thank you for remind. I will soon upload again.

meier-rene commented 5 years ago

Well done. Thank you for your contribution.