Ab-origin: an enhanced tool to identify the sourcing gene segments in germline for rearranged antibodies

BMC Bioinformatics

Table 2 Results of a set of 1000 simulated sequences with different mutation rates.

	Number of being rejected^a	IGHD		IGHD			IGHJ
		Wrong pickups^b	Accuracy (%)^c	Missing^d	Wrong pickups	Accuracy (%)	Wrong pickups	Accuracy (%)
Ab-origin	0	30	97.0	7	113	88.0	19	98.1
IMGT\V-QUEST	0	30	97.0	29	147	82.4*	44	95.6*
SoDA	0	87	91.3*	3	161	83.6*	83	91.7*
JOINSOLVER	26	161	81.3*	9	214	75.1*	46	92.8*
VDJsolver	97	80	82.3*	73	86	74.4*	35	86.8*
iHMMune-align	68	67	86.5*	22	111	79.9*	37	89.5*

The numbers represent the results from six programs. A wrong pickup means that the identified gene segment was not the one we used exactly in simulation.
^a Number of the rearranged sequences cannot be deciphered by the tool due to various reasons. Some tools give the explanation for rejection, i.e. JOINSOLVER could not well define the CDR3 region [15]; iHMMune-align is not intended for gaps in IGHV and short length of IGHV [17].
^b Number of the wrong pickups among the sequences already been deciphered.
^c Accuracy = 1-(wrong pickups+ number of missing+ number of being rejected)/1000
^d Number of the sequences already been deciphered by the tool but didn't find the IGHD alignment because of respective restrictions.
* Demonstrate that the accuracy of Ab-origin is significantly higher than that of the tool.

ISSN: 1471-2105