Table 2.

Alignment and tree accuracy metrics for all methods on 1,000 sequences

					Delta-RF
Data	Method	Prec.	Rec.	TC	RAxML	FT-2
Indelible	P(Default)	95.1 %	94.6 %	4.5 %	1.86 %	0.68 %
M2	P+BAli-Phy	98.7 %	98.7 %	14.6 %	0.29 %	-0.72 %
	P+MAFFT-L	97.2 %	97.0 %	6.8 %	0.75 %	-0.20 %
	MAFFT-L	80.2 %	75.0 %	1.4 %	15.73 %	8.74 %
	MAFFT-def	1.0 %	0.4 %	0.0 %	(not run)	(not run)
RNAsim	P(default)	90.3 %	90.4 %	3.5 %	0.56 %	0.33 %
	P+BAli-Phy	92.1 %	92.1 %	8.5 %	0.70 %	0.42 %
	P+MAFFT-L	88.8 %	89.0 %	3.9 %	0.34 %	0.45 %
	MAFFT-L	91.8 %	91.5 %	2.9 %	0.73 %	6.47 %
	MAFFT-def	83.7 %	71.5 %	1.4 %	(not run)	(not run)
Rose L1	P(default)	90.9 %	90.6 %	15.9 %	2.07 %	2.24 %
	P+BAli-Phy	91.8 %	91.7 %	33.2 %	1.47 %	1.51 %
	P+MAFFT-L	90.0 %	89.8 %	21.8 %	1.98 %	2.00 %
	MAFFT-L	84.1 %	76.6 %	6.4 %	3.45 %	3.15 %
	MAFFT-def	1.1 %	0.4 %	0.0 %	(not run)	(not run)
Rose M1	P(default)	79.7 %	79.0 %	9.0 %	5.35 %	6.26 %
	P+BAli-Phy	79.8 %	79.6 %	24.4 %	4.70 %	5.45 %
	P+MAFFT-L	78.6 %	78.2 %	12.9 %	5.96 %	5.89 %
	MAFFT-L	74.9 %	63.3 %	3.0 %	3.64 %	3.90 %
	MAFFT-def	1.2 %	0.5 %	0.0 %	(not run)	(not run)
Rose S1	P(default)	85.3 %	85.1 %	2.8 %	3.94 %	4.29 %
	P+BAli-Phy	84.3 %	84.3 %	10.3 %	2.26 %	3.59 %
	P+MAFFT-L	83.5 %	83.3 %	4.8 %	3.55 %	4.38 %
	MAFFT-L	76.2 %	68.2 %	0.5 %	3.80 %	3.79 %
	MAFFT-def	1.2 %	0.5 %	0.0 %	(not run)	(not run)

Note that precision, recall and TC are accuracy metrics (so larger is better) but Delta-RF is an error metric (so smaller is better). Metrics are averages over 10 replicates. Method names have been shortened slightly for space: P(default) refers to PASTA(default), P+(...) is shorthand for PASTA+(...), MAFFT-def refers to default MAFFT, and MAFFT-L refers to MAFFT L-INS-i. Bold numbers indicate best performing method