Table 3.

Statistics of identified and reliably measured transcripts, adding genes identified de novo.

	Bowtie + MapAl + Cufflinks + models		TopHat + Cufflinks (de novo)		Together (models + de novo)
	Identified	Reliable	Identified	Reliable	Identified	Reliable
Set 1	101,169	56,980	164,138	11,288	265,307	68,268
Set 2	105,091	21,091	367,383	3,579	472,474	24,670

The left-most column of  Table 2 includes results for both known transcripts and genes identified de novo. Restricting the pipeline to known splice forms and making use of known splice form sequences as much as possible has considerably increased the number of transcripts that can be assessed reliably. These statistics are shown again in the left-most column of this table. We can now consider adding genes identified only de novo. Statistics for these genes are shown in the middle column. Taken together, we can achieve the most powerful analysis, namely combining the output of the [Bowtie + MapAl + Cufflinks + models] pipeline with the additional results from [TopHat + Cufflinks (de novo)].