Figure 6

Library classification procedure. A triplet of fields consisting of tissue, organ, and title for each EST library from the dbEST report files was extracted. For triplets with records containing the same description under both tissue and organ or no record for one, but not both, the library can be automatically classified into the category described in the tissue/organ field. This automatic assignment was then checked with title to avoid omission of useful information, such as disease state, that could be hidden in the title fields. In some cases (case 2 and 4), the tissue and organ fields were different or both were empty. If different, the library was temporarily classified according to tissue, then modified manually if the other two fields contained more information that could be used for the modification. If both were null, the classification relied on the title.