IV. Quality control
The tomato EST sequencing project uses libraries from various sources (mainly Cornell) which have been arrayed and sequenced at TiGR; CUGI serves as a distribution hub. We purchased the 202 plates that contained clones representing the ~4,300 sequences of ToRuGs from CUGI as bacterial glycerol master stocks in 384 well plates.
Preliminary sampling showed that ~9.6% (sample size = 3,072 wells) of the clones in the glycerol masters from CUGI were not turbid, and 2.6% of the clones seemed “abnormal” (e.g., very cloudy, precipitates, etc.). Manual and robotic re-arraying confirmed that “clear glycerols” failed to contain recoverable colonies, and the "abnormal" wells appear to represent contaminated cultures. Neither of these types of failure show obvious patterns.
Using a Q-bot robotic picker, the 384-clones defining ToRuGs plate I were arrayed from 24 of the plates obtained from CUGI. Plasmid DNA was extracted and sequences determined for 96 of the clones, and compared to LeGI at TiGR:
- ~10% of the sequences matched a TC in the LeGI, but was the wrong TC for that clone address. In one case, the sequence matched an adjacent clone.
- ~20% of the sequences did not match highly to any TiGR TC, or were low quality sequence, possibly reflecting contamination of the glycerol master plates.
