How are the sequence data processed and what is the quality of the sequence?
CK Library: The sequences were trimmed and edited manually using Sequencher 3.1 software. When the 5′ and 3′ sequences overlapped, contigs were constructed. All submitted sequences have an error rate of 3% or less. LD, HL, GM, GH and LP Libraries: BDGP EST lengths average 546 bp, with high quality sequence averaging 447 bp. LD, HL, GM, GH and LP Libraries: The sequences are first processed by trimming the vector sequences (about 60 bp) from their 5′ ends. The 3′ end of the sequence is trimmed based on two quality cutoffs. The first cutoff marks the end of the so-called “high quality” sequence. We are using identical criteria to those being used in the HHMI/WashU Mouse EST Project, kindly provided by L. Hillier. By comparing our EST sequences to those in the database in cases where the corresponding gene has been sequenced, we estimate that the “high quality” sequence is more than 99% accurate. Finally, we trim the sequence at a point where we estimate the accuracy falls below 97%. The
