Apollo: A Sequencing-Technology-Independent, Scalable, and Accurate   Assembly Polishing Algorithm

Can Firtina; Jeremie S. Kim; Mohammed Alser; Damla Senol Cali; A.; Ercument Cicek; Can Alkan; Onur Mutlu

arXiv:1902.04341·q-bio.GN·October 29, 2020

Apollo: A Sequencing-Technology-Independent, Scalable, and Accurate Assembly Polishing Algorithm

Can Firtina, Jeremie S. Kim, Mohammed Alser, Damla Senol Cali, A., Ercument Cicek, Can Alkan, Onur Mutlu

PDF

1 Repo

TL;DR

Apollo is a universal, scalable assembly polishing algorithm that effectively integrates reads from all sequencing technologies to improve genome assembly accuracy without size or technology limitations.

Contribution

It introduces a novel pHMM-based approach that unifies polishing across different sequencing technologies and scales to large genomes in a single run.

Findings

01

Apollo uses reads from all technologies in one run

02

It scales well to large genome assemblies

03

It outperforms existing algorithms in accuracy and scalability

Abstract

Long reads produced by third-generation sequencing technologies are used to construct an assembly (i.e., the subject's genome), which is further used in downstream genome analysis. Unfortunately, long reads have high sequencing error rates and a large proportion of bps in these long reads are incorrectly identified. These errors propagate to the assembly and affect the accuracy of genome analysis. Assembly polishing algorithms minimize such error propagation by polishing or fixing errors in the assembly by using information from alignments between reads and the assembly (i.e., read-to-assembly alignment information). However, assembly polishing algorithms can only polish an assembly using reads either from a certain sequencing technology or from a small assembly. Such technology-dependency and assembly-size dependency require researchers to 1) run multiple polishing algorithms and 2)…

Tables22

Table 1. Table S1: Comparison between using a hybrid set of reads with Apollo and running other polishing tools multiple times to polish a Canu-generated assembly

Dataset	First Run	Second Run	Aligned	Accuracy	Polishing	Runtime	Memory
			Bases (%)		Score		(GB)
E. coli O157	—	—	99.94	0.9998	0.9992	43m 53s	3.79
E. coli O157	Apollo (Hybrid)	—	99.94	0.9999	0.9993	8h 16m 08s	13.85
E. coli O157	Racon (PacBio)	Racon (Illumina)	99.94	0.9994	0.9988	21m 44s	22.65
E. coli O157	Racon (PacBio)	Racon (PacBio)	99.94	0.9984	0.9978	4m 58s	2.43
E. coli O157	Pilon (Illumina)	Pilon (Illumina)	99.94	0.9999	0.9993	4m 10s	11.40
E. coli O157	Pilon (Illumina)	Racon (PacBio)	99.94	0.9986	0.9980	4m 58s	11.40
E. coli O157	Quiver (PacBio)	Quiver (Pacbio)	99.94	0.9998	0.9992	13m 06s	1.98
E. coli O157	Quiver (PacBio)	Pilon (Illumina)	99.94	0.9998	0.9992	5m 01s	7.50
E. coli O157	Quiver (PacBio)	Racon (PacBio)	99.94	0.9986	0.9980	5m 13s	2.48
E. coli O157:H7	—	—	100	0.9998	0.9998	43m 19s	3.39
E. coli O157:H7	Apollo (Hybrid)	—	100	0.9999	0.9999	5h 58m 05s	8.86
E. coli O157:H7	Racon (PacBio)	Racon (Illumina)	100	0.9995	0.9995	9m 43s	6.56
E. coli O157:H7	Racon (PacBio)	Racon (PacBio)	100	0.9970	0.9970	5m 36s	2.24
E. coli O157:H7	Pilon (Illumina)	Pilon (Illumina)	100	0.9998	0.9998	35m 12s	10.79
E. coli O157:H7	Pilon (Illumina)	Racon (PacBio)	100	0.9996	0.9996	6m 04s	10.75
E. coli K-12	—	—	99.98	0.9794	0.9792	34h 21m 46s	5.06
E. coli K-12	Apollo (Hybrid)	—	99.99	0.9953	0.9952	9h 09m 50s	9.35
E. coli K-12	Racon (ONT)	Racon (Illumina)	100	0.9996	0.9996	11m 05s	5.10
E. coli K-12	Racon (ONT)	Racon (ONT)	100	0.9851	0.9851	14m 45s	4.20
E. coli K-12	Pilon (Illumina)	Pilon (Illumina)	99.99	0.9993	0.9992	18m 55s	8.84
E. coli K-12	Pilon (Illumina)	Racon (ONT)	99.99	0.9997	0.9996	15m 51s	8.84
E. coli K-12	Nanopolish (ONT)	Nanopolish (ONT)	99.98	0.9929	0.9927	25h 39m 17s	4.84
E. coli K-12	Nanopolish (ONT)	Pilon (Illumina)	99.99	0.9992	0.9991	9h 45m 01s	18.10
E. coli K-12	Nanopolish (ONT)	Racon (ONT)	100	0.9866	0.9866	9h 42m 24s	4.54
Yeast S288C	—	—	99.89	0.9998	0.9987	1h 20m 39s	6.24
Yeast S288C	Apollo (Hybrid)	—	99.89	0.9998	0.9987	11h 08m 41s	6.38
Yeast S288C	Racon (PacBio)	Racon (Illumina)	99.89	0.9994	0.9983	38m 21s	6.93
Yeast S288C	Racon (PacBio)	Racon (PacBio)	99.89	0.9949	0.9938	49m 52s	6.93
Yeast S288C	Pilon (Illumina)	Pilon (Illumina)	99.89	0.9998	0.9987	1m 10s	11.85
Yeast S288C	Pilon (Illumina)	Racon (PacBio)	99.89	0.9960	0.9949	21m 42s	11.85
Yeast S288C	Quiver (PacBio)	Quiver (Pacbio)	98.95	0.9998	0.9893	23m 23s	2.96
Yeast S288C	Quiver (PacBio)	Pilon (Illumina)	98.95	0.9998	0.9893	12m 47s	13.28
Yeast S288C	Quiver (PacBio)	Racon (PacBio)	98.93	0.9968	0.9861	40m 04s	6.69

Table 2. Table S2: Comparison between using a hybrid set of reads with Apollo and running other polishing tools multiple times to polish a Miniasm-generated assembly

Dataset	First Run	Second Run	Aligned	Accuracy	Polishing	Runtime	Memory
			Bases (%)		Score		(GB)
E. coli O157	—	—	94.93	0.9000	0.8544	1m 48s	10.03
E. coli O157	Apollo (Hybrid)	—	98.70	0.9866	0.9738	3h 51m 51s	12.08
E. coli O157	Racon (PacBio)	Racon (Illumina)	99.37	0.9992	0.9929	21m 19s	22.66
E. coli O157	Racon (PacBio)	Racon (PacBio)	99.51	0.9980	0.9931	5m 00s	2.46
E. coli O157	Pilon (Illumina)	Pilon (Illumina)	96.88	0.9872	0.9564	34m 53s	18.60
E. coli O157	Pilon (Illumina)	Racon (PacBio)	98.87	0.9970	0.9857	35m 26s	18.60
E. coli O157	Quiver (PacBio)	Quiver (PacBio)	99.85	0.9994	0.9979	13m 45s	5.05
E. coli O157	Quiver (PacBio)	Pilon (Illumina)	99.80	0.9994	0.9974	9m 42s	4.76
E. coli O157	Quiver (PacBio)	Racon (PacBio)	99.81	0.9984	0.9965	10m 29s	2.49
E. coli O157:H7	—	—	88.56	0.8798	0.7792	2m 57s	6.27
E. coli O157:H7	Apollo (Hybrid)	—	97.53	0.9804	0.9562	2h 54m 55s	8.34
E. coli O157:H7	Racon (PacBio)	Racon (Illumina)	99.02	0.9991	0.9893	9m 24s	6.56
E. coli O157:H7	Racon (PacBio)	Racon (PacBio)	99.22	0.9954	0.9876	5m 31s	2.24
E. coli O157:H7	Racon (PacBio)	Pilon (Illumina)	99.12	0.9981	0.9893	20m 37s	12.57
E. coli O157:H7	Pilon (Illumina)	Pilon (Illumina)	96.32	0.9896	0.9532	35m 12s	15.84
E. coli K-12	—	—	86.68	0.8503	0.7370	4m 04s	16.47
E. coli K-12	Apollo (Hybrid)	—	97.53	0.9419	0.9186	2h 18m 33s	9.12
E. coli K-12	Racon (ONT)	Racon (Illumina)	99.51	0.9992	0.9943	8m 38s	5.17
E. coli K-12	Racon (ONT)	Racon (ONT)	99.78	0.9840	0.9818	11m 43s	4.06
E. coli K-12	Pilon (Illumina)	Pilon (Illumina)	89.61	0.9622	0.8622	32m 03s	17.78
E. coli K-12	Pilon (Illumina)	Racon (ONT)	99.43	0.9979	0.9922	25m 15s	32.15
E. coli K-12	Nanopolish (ONT)	Nanopolish (ONT)	97.35	0.9488	0.9236	241h 56m 10s	8.49
E. coli K-12	Nanopolish (ONT)	Pilon (Illumina)	96.48	0.9769	0.9425	117h 29m 47s	32.15
E. coli K-12	Nanopolish (ONT)	Racon (ONT)	99.62	0.9814	0.9776	117h 08m 16s	8.49
Yeast S288C	—	—	95.05	0.8923	0.8481	2m 20s	16.59
Yeast S288C	Apollo (Hybrid)	—	98.49	0.9709	0.9562	6h 37m 46s	5.96
Yeast S288C	Racon (PacBio)	Racon (Illumina)	99.26	0.9986	0.9912	23m 51s	6.75
Yeast S288C	Racon (PacBio)	Racon (PacBio)	99.33	0.9937	0.9879	43m 00s	6.75
Yeast S288C	Racon (PacBio)	Pilon (Illumina)	99.23	0.9977	0.9900	22m 07s	14.86
Yeast S288C	Pilon (Illumina)	Pilon (Illumina)	95.80	0.9595	0.9192	2m 35s	15.31
Yeast S288C	Quiver (PacBio)	Quiver (PacBio)	99.42	0.9997	0.9939	24m 49s	4.14
Yeast S288C	Quiver (PacBio)	Pilon (Illumina)	99.45	0.9996	0.9941	12m 23s	13.40
Yeast S288C	Quiver (PacBio)	Racon (PacBio)	99.50	0.9965	0.9915	29m 31s	6.39

Table 3. Table S3: Assembly polishing performance of the tools for the E. coli O157 dataset

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	Aligned	Accuracy	Polishing	Runtime	Memory
			of the Reads	Algorithm	Bases (%)		Score		(GB)
PacBio	Miniasm	—	—	—	94.93	0.9000	0.8544	1m 48s	10.03
PacBio	Miniasm	Minimap2	PacBio	Apollo	98.49	0.9798	0.9650	2h 27m 49s	7.07
PacBio	Miniasm	Minimap2	PacBio	Pilon	96.43	0.9528	0.9188	1h 31m 32s	17.68
PacBio	Miniasm	Minimap2	PacBio	Racon	99.35	0.9951	0.9886	2m 13s	2.44
PacBio	Miniasm	pbalign	PacBio	Quiver	99.80	0.9993	0.9973	7m 31s	0.51
PacBio	Miniasm	Minimap2	Illumina	Apollo	97.61	0.9816	0.9581	4h 25m 17s	9.22
PacBio	Miniasm	Minimap2	Illumina	Pilon	96.52	0.9775	0.9435	32m 48s	18.60
PacBio	Miniasm	Minimap2	Illumina	Racon	96.45	0.9876	0.9525	14m 09s	21.57
PacBio	Miniasm	BWA-MEM	Illumina	Apollo	96.62	0.9738	0.9409	3h 32m 45s	9.21
PacBio	Miniasm	BWA-MEM	Illumina	Pilon	96.13	0.9693	0.9318	31m 21s	18.45
PacBio	Miniasm	BWA-MEM	Illumina	Racon	96.90	0.9813	0.9509	12m 05s	20.85
PacBio	Canu	—	—	—	99.94	0.9998	0.9992	43m 53s	3.79
PacBio	Canu	Minimap2	PacBio	Apollo	99.94	0.9997	0.9991	3h 42m 03s	8.82
PacBio	Canu	Minimap2	PacBio	Racon	99.94	0.9986	0.9980	2m 17s	2.34
PacBio	Canu	pbalign	PacBio	Quiver	99.94	0.9998	0.9992	7m 06s	0.20
PacBio	Canu	BWA-MEM	Illumina	Apollo	99.94	0.9999	0.9993	4h 49m 15s	11.05
PacBio	Canu	BWA-MEM	Illumina	Pilon	99.94	0.9998	0.9992	2m 05s	11.40
PacBio	Canu	BWA-MEM	Illumina	Racon	99.94	0.9999	0.9993	14m 58s	21.04
PacBio (30 $\times$ )	Miniasm^∗	—	—	—	—	—	—	—	—
PacBio (30 $\times$ )	Canu	—	—	—	99.98	0.9981	0.9979	21m 03s	3.70
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Apollo	99.98	0.9982	0.9980	43m 32s	8.00
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Racon	99.98	0.9980	0.9978	15s	0.59
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Apollo	99.97	0.9976	0.9973	46m 10s	7.99
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Racon	99.98	0.9983	0.9981	7s	0.37
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Apollo	99.98	0.9997	0.9995	4h 48m 31s	10.35
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Pilon	99.98	0.9998	0.9996	3m 03s	8.52
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Racon	99.98	0.9997	0.9995	14m 42s	21.04

Table 4. Table S4: K-mer similarity between Illumina reads and the E. coli O157 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	11-mer	21-mer	31-mer	51-mer
			of the Reads	Algorithm	Sim. (%)	Sim. (%)	Sim. (%)	Sim. (%)
PacBio	Reference	—	—	—	100 / 100	99.89 / 99.98	99.92 / 99.96	99.66 / 99.96
PacBio	Miniasm	—	—	—	90.67 / 83.48	14.31 / 13.53	5.61 / 5.21	1.12 / 1.04
PacBio	Miniasm	Minimap2	PacBio	Apollo	96.19 / 94.94	76.20 / 74.70	66.76 / 64.01	54.77 / 52.38
PacBio	Miniasm	Minimap2	PacBio	Pilon	93.63 / 89.91	46.18 / 44.24	31.07 / 28.92	14.57 / 13.70
PacBio	Miniasm	Minimap2	PacBio	Racon	99.47 / 98.70	94.89 / 94.11	91.11 / 89.05	85.22 / 84.67
PacBio	Miniasm	pbalign	PacBio	Quiver	100 / 99.61	99.81 / 99.06	99.65 / 98.41	99.16 / 98.31
PacBio	Miniasm	Minimap2	Illumina	Apollo	97.11 / 95.42	83.33 / 82.33	78.23 / 76.56	71.05 / 69.02
PacBio	Miniasm	Minimap2	Illumina	Pilon	96.52 / 93.93	83.74 / 80.15	82.25 / 77.44	79.02 / 74.49
PacBio	Miniasm	Minimap2	Illumina	Racon	97.31 / 96.42	90.35 / 90.02	88.61 / 87.88	87.98 / 87.34
PacBio	Miniasm	BWA-MEM	Illumina	Apollo	96.98 / 94.19	80.06 / 77.20	75.18 / 72.08	67.71 / 64.42
PacBio	Miniasm	BWA-MEM	Illumina	Pilon	96.32 / 93.20	79.65 / 75.30	76.75 / 72.32	72.92 / 67.16
PacBio	Miniasm	BWA-MEM	Illumina	Racon	96.91 / 95.10	85.89 / 85.27	84.00 / 83.88	82.36 / 81.06
PacBio	Canu	—	—	—	100 / 99.93	99.63 / 99.78	99.46 / 99.42	98.93 / 99.00
PacBio	Canu	Minimap2	PacBio	Apollo	100 / 99.93	99.50 / 99.74	99.17 / 99.50	98.50 / 99.11
PacBio	Canu	Minimap2	PacBio	Racon	99.87 / 99.74	98.44 / 98.52	97.37 / 97.39	95.63 / 95.78
PacBio	Canu	pbalign	PacBio	Quiver	100 / 100	99.80 / 99.72	99.67 / 99.44	99.40 / 99.25
PacBio	Canu	BWA-MEM	Illumina	Apollo	100 / 100	99.83 / 99.91	99.73 / 99.77	99.59 / 99.61
PacBio	Canu	BWA-MEM	Illumina	Pilon	100 / 100	99.83 / 99.93	99.73 / 99.77	99.59 / 99.62
PacBio	Canu	BWA-MEM	Illumina	Racon	100 / 100	99.81 / 99.91	99.71 / 99.75	99.57 / 99.53
PacBio (30 $\times$ )	Canu	—	—	—	99.47 / 99.41	96.74 / 96.88	95.20 / 94.92	92.31 / 91.39
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Apollo	99.61 / 99.41	97.04 / 97.40	95.41 / 95.63	92.67 / 92.48
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Racon	99.80 / 99.61	97.00 / 97.34	95.12 / 95.16	92.63 / 92.98
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Apollo	99.41 / 99.41	97.00 / 97.47	95.31 / 95.72	92.44 / 92.93
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Racon	99.67 / 99.48	97.48 / 98.19	96.00 / 96.56	93.12 / 94.07
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Apollo	100 / 99.93	99.83 / 99.54	99.69 / 99.52	99.55 / 99.23
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Pilon	100 / 99.93	99.83 / 99.70	99.69 / 99.58	99.51 / 99.31
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Racon	100 / 99.93	99.89 / 99.63	99.73 / 99.62	99.55 / 99.29

Table 5. Table S5: Quality assessment of the E. coli O157 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	GC	Mapped	Properly	Avg.	Coverage
			of the Reads	Algorithm	(%)	Reads (%)	Paired (%)	Coverage	$\geq$ 10 $\times$ (%)
PacBio	Reference	—	—	—	50.48	99.92	99.49	564	99.94
PacBio	Miniasm	—	—	—	49.88	92.08	87.50	434	87.90
PacBio	Miniasm	Minimap2	PacBio	Apollo	50.28	98.74	97.43	531	96.19
PacBio	Miniasm	Minimap2	PacBio	Pilon	50.14	99.17	97.20	526	93.78
PacBio	Miniasm	Minimap2	PacBio	Racon	50.52	99.63	99.03	542	98.35
PacBio	Miniasm	pbalign	PacBio	Quiver	50.56	99.83	99.40	545	98.56
PacBio	Miniasm	Minimap2	Illumina	Apollo	50.37	96.49	94.60	513	93.74
PacBio	Miniasm	Minimap2	Illumina	Pilon	50.36	95.58	92.04	499	89.57
PacBio	Miniasm	Minimap2	Illumina	Racon	50.45	96.48	94.73	514	94.11
PacBio	Miniasm	BWA-MEM	Illumina	Apollo	50.30	95.55	92.22	498	89.58
PacBio	Miniasm	BWA-MEM	Illumina	Pilon	50.30	94.48	89.64	478	86.54
PacBio	Miniasm	BWA-MEM	Illumina	Racon	50.37	94.63	90.69	508	90.76
PacBio	Canu	—	—	—	50.36	99.90	99.46	547	99.73
PacBio	Canu	Minimap2	PacBio	Apollo	50.36	99.90	99.46	547	99.92
PacBio	Canu	Minimap2	PacBio	Racon	50.35	99.89	99.44	547	99.89
PacBio	Canu	pbalign	PacBio	Quiver	50.36	99.90	99.46	547	99.38
PacBio	Canu	BWA-MEM	Illumina	Apollo	50.36	99.90	99.46	547	99.73
PacBio	Canu	BWA-MEM	Illumina	Pilon	50.36	99.90	99.46	547	99.73
PacBio	Canu	BWA-MEM	Illumina	Racon	50.36	99.90	99.46	547	99.73
PacBio (30 $\times$ )	Canu	—	—	—	50.44	99.89	99.42	560	99.61
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Apollo	50.46	99.89	99.44	560	99.91
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	Racon	50.44	99.89	99.43	560	99.94
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Apollo	50.46	99.89	99.42	560	99.92
PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr.)	Racon	50.46	99.89	99.42	560	99.97
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Apollo	50.47	99.89	99.44	560	99.70
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Pilon	50.47	99.89	99.44	560	99.71
PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	Racon	50.47	99.89	99.43	560	99.69

Table 6. Table S6: Assembly polishing performance of the tools for the E. coli O157:H7 dataset

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	Aligned	Accuracy	Polishing	Runtime	Memory
			of the Reads	Algorithm	Bases (%)		Score		(GB)
PacBio	Miniasm	—	—	—	88.56	0.8798	0.7792	2m 57s	6.27
PacBio	Miniasm	Minimap2	PacBio	Apollo	96.99	0.9636	0.9346	1h 10m 23s	7.07
PacBio	Miniasm	Minimap2	PacBio	Racon	98.94	0.9899	0.9794	2m 24s	2.14
PacBio	Miniasm	Minimap2	Illumina	Apollo	96.06	0.9781	0.9396	2h 17m 28s	5.66
PacBio	Miniasm	Minimap2	Illumina	Pilon	95.09	0.9791	0.9310	28m 54s	15.84
PacBio	Miniasm	Minimap2	Illumina	Racon	96.17	0.9883	0.9504	4m 39s	6.29
PacBio	Canu	—	—	—	100	0.9998	0.9998	43m 19s	3.39
PacBio	Canu	Minimap2	PacBio	Apollo	100	0.9997	0.9997	2h 57m 18s	7.58
PacBio	Canu	Minimap2	PacBio	Racon	100	0.9975	0.9975	2m 50s	2.23
PacBio	Canu	Minimap2	Illumina	Apollo	100	0.9997	0.9997	3h 10m 16s	6.18
PacBio	Canu	Minimap2	Illumina	Pilon	100	0.9999	0.9999	1m 27s	10.75
PacBio	Canu	Minimap2	Illumina	Racon	100	0.9996	0.9996	7m 14s	6.53

Table 7. Table S7: K-mer similarity between Illumina reads and the E. coli O157:H7 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	11-mer	21-mer	31-mer	51-mer
			of the Reads	Algorithm	Sim. (%)	Sim. (%)	Sim. (%)	Sim. (%)
E. coli O157:H7	Reference	—	—	—	99.93 / 100	99.78 / 99.94	99.73 / 99.96	99.70 / 99.92
E. coli O157:H7	Miniasm	—	—	—	91.14 / 81.04	9.01 / 7.94	3.25 / 2.74	0.37 / 0.33
E. coli O157:H7	Miniasm	Minimap2	PacBio	Apollo	96.46 / 91.36	61.52 / 57.92	52.73 / 48.27	35.22 / 32.38
E. coli O157:H7	Miniasm	Minimap2	PacBio	Racon	98.10 / 96.95	88.45 / 85.70	84.37 / 80.22	74.61 / 70.87
E. coli O157:H7	Miniasm	Minimap2	Illumina	Apollo	97.97 / 93.43	81.92 / 78.79	77.05 / 72.69	66.69 / 63.21
E. coli O157:H7	Miniasm	Minimap2	Illumina	Pilon	97.64 / 92.25	85.57 / 79.87	84.74 / 78.02	80.92 / 75.85
E. coli O157:H7	Miniasm	Minimap2	Illumina	Racon	98.36 / 94.57	91.28 / 89.04	90.77 / 87.49	88.78 / 87.23
E. coli O157:H7	Canu	—	—	—	99.80 / 99.93	99.41 / 99.57	99.13 / 99.46	99.06 / 98.99
E. coli O157:H7	Canu	Minimap2	PacBio	Apollo	99.80 / 99.93	99.35 / 99.57	99.08 / 99.44	98.82 / 98.88
E. coli O157:H7	Canu	Minimap2	PacBio	Racon	99.54 / 99.61	96.81 / 96.67	95.27 / 95.22	91.84 / 91.72
E. coli O157:H7	Canu	Minimap2	Illumina	Apollo	99.93 / 99.93	99.48 / 99.85	99.17 / 99.71	98.95 / 99.70
E. coli O157:H7	Canu	Minimap2	Illumina	Pilon	99.87 / 100	99.78 / 99.91	99.73 / 99.88	99.70 / 99.79
E. coli O157:H7	Canu	Minimap2	Illumina	Racon	99.80 / 100	99.31 / 99.83	99.00 / 99.88	98.63 / 99.47

Table 8. Table S8: Quality assessment of the E. coli O157:H7 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	GC	Mapped	Properly	Avg.	Coverage
			of the Reads	Algorithm	(%)	Reads (%)	Paired (%)	Coverage	$\geq$ 10 $\times$ (%)
E. coli O157:H7	Reference	—	—	—	50.43	97.42	94.3	183	99.93
E. coli O157:H7	Miniasm	—	—	—	49.61	80.51	68.24	108	76.01
E. coli O157:H7	Miniasm	Minimap2	PacBio	Apollo	50.09	95.0	88.69	163	91.74
E. coli O157:H7	Miniasm	Minimap2	PacBio	Racon	50.55	97.03	93.06	173	96.59
E. coli O157:H7	Miniasm	Minimap2	Illumina	Apollo	50.39	93.6	87.69	162	90.65
E. coli O157:H7	Miniasm	Minimap2	Illumina	Pilon	50.36	93.01	85.66	159	86.75
E. coli O157:H7	Miniasm	Minimap2	Illumina	Racon	50.48	93.84	88.52	163	91.67
E. coli O157:H7	Canu	—	—	—	50.43	97.42	94.32	182	99.71
E. coli O157:H7	Canu	Minimap2	PacBio	Apollo	50.44	97.42	94.32	182	99.87
E. coli O157:H7	Canu	Minimap2	PacBio	Racon	50.41	97.4	94.22	182	99.73
E. coli O157:H7	Canu	Minimap2	Illumina	Apollo	50.45	97.42	94.31	182	99.95
E. coli O157:H7	Canu	Minimap2	Illumina	Pilon	50.44	97.42	94.33	182	99.71
E. coli O157:H7	Canu	Minimap2	Illumina	Racon	50.45	97.42	94.29	182	99.98

Table 9. Table S9: Assembly polishing performance of the tools for E. coli K-12 MG1655 dataset

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	Aligned	Accuracy	Polishing	Runtime	Memory
			of the Reads	Algorithm	Bases (%)		Score		(GB)
ONT	Miniasm	—	—	—	86.68	0.8503	0.7370	4m 04s	16.47
ONT	Miniasm	Minimap2	ONT	Apollo	97.50	0.9209	0.8979	1h 40m 08s	7.96
ONT	Miniasm	Minimap2	ONT	Nanopolish	96.01	0.9182	0.8816	117h 02m 10s	8.49
ONT	Miniasm	Minimap2	ONT	Racon	99.41	0.9769	0.9711	4m 55s	3.70
ONT	Miniasm	Minimap2	Illumina	Apollo	89.41	0.9291	0.8307	54m 46s	6.20
ONT	Miniasm	Minimap2	Illumina	Pilon	89.22	0.9310	0.8306	17m 28s	10.58
ONT	Canu	—	—	—	99.98	0.9794	0.9792	34h 21m 46s	5.06
ONT	Canu	Minimap2	ONT	Apollo	99.99	0.9803	0.9802	6h 08m 05s	8.09
ONT	Canu	Minimap2	ONT	Nanopolish	99.98	0.9925	0.9923	9h 35m 26s	4.54
ONT	Canu	Minimap2	ONT	Racon	100	0.9840	0.9840	7m 22s	4.20
ONT	Canu	Minimap2	Illumina	Apollo	99.96	0.9982	0.9978	2h 09m 47s	6.43
ONT	Canu	Minimap2	Illumina	Pilon	99.99	0.9987	0.9986	11m 59s	8.84
ONT (30 $\times$ )	Miniasm^∗	—	—	—	—	—	—	—	—
ONT (30 $\times$ )	Canu	—	—	—	99.98	0.9744	0.9742	3h 17m 47s	4.54
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Apollo	99.98	0.9752	0.9750	40m 37s	7.74
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Nanopolish	99.99	0.9857	0.9856	4h 07m 06s	2.15
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Racon	100	0.9825	0.9825	20s	0.59
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Apollo	99.96	0.9755	0.9751	46m 40s	7.75
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Racon	100	0.9799	0.9799	9s	0.42

Table 10. Table S10: K-mer similarity between Illumina reads and the E. coli K-12 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	11-mer	21-mer	31-mer	51-mer
			of the Reads	Algorithm	Sim. (%)	Sim. (%)	Sim. (%)	Sim. (%)
ONT	Reference	—	—	—	99.79 / 100	99.37 / 99.70	99.35 / 99.51	99.22 / 99.65
ONT	Miniasm	—	—	—	82.92 / 80.97	13.49 / 14.57	5.40 / 5.59	1.22 / 1.29
ONT	Miniasm	Minimap2	ONT	Apollo	88.09 / 87.01	39.46 / 41.06	26.10 / 27.06	12.11 / 12.20
ONT	Miniasm	Minimap2	ONT	Nanopolish	89.67 / 87.09	47.47 / 48.79	38.19 / 37.81	25.04 / 25.30
ONT	Miniasm	Minimap2	ONT	Racon	93.25 / 95.02	75.24 / 74.16	63.69 / 63.36	48.72 / 47.87
ONT	Miniasm	Minimap2	Illumina	Apollo	91.25 / 87.17	50.96 / 53.20	44.37 / 44.28	32.54 / 32.75
ONT	Miniasm	Minimap2	Illumina	Pilon	89.60 / 86.45	56.27 / 58.38	51.30 / 52.20	44.28 / 45.29
ONT	Canu	—	—	—	92.08 / 95.91	76.08 / 76.15	66.05 / 66.09	49.94 / 49.87
ONT	Canu	Minimap2	ONT	Apollo	92.15 / 95.91	76.93 / 77.04	67.52 / 67.37	51.35 / 50.94
ONT	Canu	Minimap2	ONT	Nanopolish	97.04 / 98.60	90.74 / 91.32	86.95 / 86.20	79.33 / 78.49
ONT	Canu	Minimap2	ONT	Racon	94.49 / 96.89	80.33 / 80.42	72.03 / 71.58	57.39 / 56.79
ONT	Canu	Minimap2	Illumina	Apollo	99.24 / 99.65	97.72 / 97.88	97.35 / 96.94	96.26 / 95.82
ONT	Canu	Minimap2	Illumina	Pilon	99.59 / 99.59	98.10 / 98.39	98.37 / 97.70	97.06 / 96.46
ONT (30 $\times$ )	Canu	—	—	—	90.36 / 94.87	71.60 / 72.15	59.89 / 59.83	41.94 / 42.42
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Apollo	91.05 / 94.84	72.62 / 73.06	60.96 / 61.17	43.50 / 43.84
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Nanopolish	95.94 / 96.80	83.00 / 82.30	75.37 / 73.76	61.85 / 60.96
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Racon	93.73 / 96.46	79.13 / 78.91	68.55 / 68.62	53.17 / 53.00
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Apollo	91.05 / 94.97	72.64 / 73.56	61.21 / 62.11	42.89 / 43.54
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Racon	92.08 / 95.91	74.79 / 76.08	64.47 / 65.09	47.06 / 47.38

Table 11. Table S11: Quality assessment of the E. coli K-12 assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	GC	Mapped	Properly	Avg.	Coverage
			of the Reads	Algorithm	(%)	Reads (%)	Paired (%)	Coverage	$\geq$ 10 $\times$ (%)
ONT	Reference	—	—	—	50.79	99.70	98.96	237	99.55
ONT	Miniasm	—	—	—	52.62	90.85	82.50	147	75.72
ONT	Miniasm	Minimap2	ONT	Apollo	52.23	97.44	94.28	216	94.84
ONT	Miniasm	Minimap2	ONT	Nanopolish	52.10	96.97	90.32	200	90.35
ONT	Miniasm	Minimap2	ONT	Racon	51.12	99.09	97.71	234	98.51
ONT	Miniasm	Minimap2	Illumina	Apollo	51.89	92.90	86.52	181	80.33
ONT	Miniasm	Minimap2	Illumina	Pilon	52.11	92.59	85.77	175	78.64
ONT	Canu	—	—	—	51.05	99.61	98.71	233	98.75
ONT	Canu	Minimap2	ONT	Apollo	50.90	99.67	98.57	234	98.31
ONT	Canu	Minimap2	ONT	Nanopolish	51.04	99.66	98.83	234	98.77
ONT	Canu	Minimap2	ONT	Racon	51.01	99.65	98.75	234	99.24
ONT	Canu	Minimap2	Illumina	Apollo	50.81	99.68	98.80	235	98.58
ONT	Canu	Minimap2	Illumina	Pilon	50.80	99.68	98.77	235	98.76
ONT (30 $\times$ )	Canu	—	—	—	51.11	99.60	98.57	234	99.04
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Apollo	51.14	99.60	98.59	234	99.19
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Nanopolish	51.12	99.65	98.72	235	98.92
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	Racon	51.05	99.64	98.78	234	99.35
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Apollo	51.14	99.60	98.65	234	99.28
ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ , Corr)	Racon	51.08	99.63	98.80	234	99.40

Table 12. Table S12: Assembly polishing performance of the tools for the Yeast S288C dataset

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	Aligned	Accuracy	Polishing	Runtime	Memory
			of the Reads	Algorithm	Bases (%)		Score		(GB)
PacBio	Miniasm	—	—	—	95.05	0.8923	0.8481	2m 23s	16.59
PacBio	Miniasm	Minimap2	PacBio	Apollo	98.44	0.9706	0.9555	6h 53m 51s	4.62
PacBio	Miniasm	Minimap2	PacBio	Racon	99.15	0.9895	0.9811	18m 55s	6.63
PacBio	Miniasm	Minimap2	PacBio	Quiver	99.44	0.9995	0.9939	16m 11s	0.26
PacBio	Miniasm	Minimap2	Illumina	Apollo	97.26	0.9733	0.9466	2h 05m 58s	2.83
PacBio	Miniasm	Minimap2	Illumina	Pilon	97.06	0.9761	0.9474	4m 00s	26.64
PacBio	Miniasm	Minimap2	Illumina	Racon	97.27	0.9835	0.9567	5m 00s	7.34
PacBio	Canu	—	—	—	99.89	0.9998	0.9987	1h 20m 39s	6.24
PacBio	Canu	Minimap2	PacBio	Apollo	98.95	0.9997	0.9892	10h 59m 10s	5.05
PacBio	Canu	Minimap2	PacBio	Racon	98.93	0.9964	0.9857	19m 16s	6.82
PacBio	Canu	Minimap2	PacBio	Quiver	98.95	0.9998	0.9893	12m 02s	0.29
PacBio	Canu	Minimap2	Illumina	Apollo	98.95	0.9998	0.9893	1h 22m 24s	3.25
PacBio	Canu	Minimap2	Illumina	Pilon	98.95	0.9998	0.9893	43s	13.83
PacBio	Canu	Minimap2	Illumina	Racon	98.95	0.9998	0.9893	2m 55s	5.15

Table 13. Table S13: K-mer similarity between Illumina reads and the Yeast S288C assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	11-mer	21-mer	31-mer	51-mer
			of the Reads	Algorithm	Sim. (%)	Sim. (%)	Sim. (%)	Sim. (%)
Yeast S288C	Reference	—	—	—	100 / 100	99.96 / 99.87	99.87 / 99.71	99.73 / 99.59
Yeast S288C	Miniasm	—	—	—	95.49 / 91.36	12.06 / 10.85	4.38 / 3.84	0.62 / 0.55
Yeast S288C	Miniasm	Minimap2	PacBio	Apollo	98.79 / 96.71	65.93 / 62.88	53.80 / 50.13	35.83 / 33.02
Yeast S288C	Miniasm	Minimap2	PacBio	Racon	99.39 / 98.63	88.15 / 86.21	82.35 / 79.89	72.60 / 69.48
Yeast S288C	Miniasm	Minimap2	PacBio	Quiver	99.89 / 99.34	99.38 / 98.42	99.07 / 98.19	98.98 / 97.63
Yeast S288C	Miniasm	Minimap2	Illumina	Apollo	98.35 / 96.65	77.96 / 74.13	69.85 / 66.35	59.06 / 55.89
Yeast S288C	Miniasm	Minimap2	Illumina	Pilon	98.84 / 96.25	84.87 / 79.60	82.25 / 77.24	80.12 / 74.60
Yeast S288C	Miniasm	Minimap2	Illumina	Racon	98.51 / 97.18	89.53 / 87.02	87.49 / 84.96	87.02 / 83.89
Yeast S288C	Canu	—	—	—	100 / 99.45	99.91 / 99.09	99.86 / 98.97	99.60 / 98.56
Yeast S288C	Canu	Minimap2	PacBio	Apollo	99.94 / 99.45	99.87 / 99.11	99.74 / 98.95	99.46 / 98.58
Yeast S288C	Canu	Minimap2	PacBio	Racon	99.94 / 99.40	96.37 / 94.96	94.20 / 92.48	89.17 / 87.70
Yeast S288C	Canu	Minimap2	PacBio	Quiver	100 / 99.62	99.93 / 99.19	99.89 / 98.95	99.76 / 98.69
Yeast S288C	Canu	Minimap2	Illumina	Apollo	100 / 99.45	99.92 / 99.10	99.88 / 98.93	99.68 / 98.58
Yeast S288C	Canu	Minimap2	Illumina	Pilon	100 / 99.45	99.94 / 99.13	99.89 / 98.95	99.74 / 98.69
Yeast S288C	Canu	Minimap2	Illumina	Racon	100 / 99.45	99.94 / 99.15	99.89 / 98.95	99.75 / 98.67

Table 14. Table S14: Quality assessment of the Yeast S288C assemblies

Dataset	Assembler	Aligner	Sequencing Tech.	Polishing	GC	Mapped	Properly	Avg.	Coverage
			of the Reads	Algorithm	(%)	Reads (%)	Paired (%)	Coverage	$\geq$ 10 $\times$ (%)
Yeast S288C	Reference	—	—	—	38.30	99.94	99.71	73	99.95
Yeast S288C	Miniasm	—	—	—	38.42	93.88	83.94	57	82.63
Yeast S288C	Miniasm	Minimap2	PacBio	Apollo	38.00	99.11	97.45	69	94.38
Yeast S288C	Miniasm	Minimap2	PacBio	Racon	38.26	99.51	98.64	70	96.34
Yeast S288C	Miniasm	Minimap2	PacBio	Quiver	38.39	99.61	99.29	71	98.04
Yeast S288C	Miniasm	Minimap2	Illumina	Apollo	38.22	97.10	94.98	66	90.53
Yeast S288C	Miniasm	Minimap2	Illumina	Pilon	38.41	96.86	88.65	66	87.78
Yeast S288C	Miniasm	Minimap2	Illumina	Racon	38.42	97.03	95.33	66	91.35
Yeast S288C	Canu	—	—	—	38.17	99.94	99.73	71	98.81
Yeast S288C	Canu	Minimap2	PacBio	Apollo	38.17	99.94	99.73	71	98.83
Yeast S288C	Canu	Minimap2	PacBio	Racon	38.09	99.94	99.23	71	98.21
Yeast S288C	Canu	Minimap2	PacBio	Quiver	38.17	99.94	99.74	71	98.74
Yeast S288C	Canu	Minimap2	Illumina	Apollo	38.17	99.94	99.73	71	98.81
Yeast S288C	Canu	Minimap2	Illumina	Pilon	38.17	99.94	99.74	71	98.81
Yeast S288C	Canu	Minimap2	Illumina	Racon	38.17	99.94	99.73	71	98.81

Table 15. Table S15: K-mer similarity between Illumina reads and the human genome assemblies

Dataset	Assembler	Aligner	Polishing	21-mer	31-mer	51-mer
			Algorithm	Sim. (%)	Sim. (%)	Sim. (%)
Human HG002	Reference	—	—	98.05 / 87.02	96.98 / 84.73	93.56 / 80.14
Human HG002	Minimap2	PacBio	Apollo	93.74 / 82.62	91.05 / 79.18	85.26 / 73.11
Human HG002	Minimap2	PacBio	Quiver^∗	94.55 / 83.49	91.50 / 79.47	84.95 / 72.36
Human HG002	Minimap2	PacBio	Racon^∗	85.96 / 74.53	79.07 / 67.58	67.19 / 56.73
Human HG002	Minimap2	Illumina	Apollo	98.33 / 87.22	97.41 / 85.05	94.26 / 80.64
Human HG002	BWA-MEM	Illumina	Apollo	98.32 / 87.17	97.39 / 84.98	94.23 / 80.57
Human HG002	BWA-MEM	Illumina	Pilon^∗	98.19 / 87.14	97.23 / 84.95	93.99 / 80.49
Human HG002	Minimap2	PacBio (9 $\times$ )	Apollo	54.00 / 43.72	45.59 / 36.91	36.82 / 30.24
Human HG002	BWA-MEM	PacBio (9 $\times$ )	Apollo	53.97 / 42.76	45.61 / 36.10	36.95 / 29.66
Human HG002	Minimap2	PacBio (9 $\times$ )	Racon	48.93 / 37.77	39.97 / 31.08	31.04 / 24.62
Human HG002	BWA-MEM	PacBio (9 $\times$ )	Racon	46.83 / 34.91	37.69 / 28.35	28.67 / 22.07

Table 16. Table S16: Quality assessment of the human genome assemblies

Dataset	Aligner	Sequencing Tech.	Polishing	GC	Mapped	Properly	Avg.	Coverage
		of the Reads	Algorithm	(%)	Reads (%)	Paired (%)	Coverage	$\geq$ 10 $\times$ (%)
Human HG002	—	—	—	40.86	99.92	98.35	10	44.82
Human HG002	Minimap2	PacBio	Apollo	40.81	99.91	97.75	10	44.81
Human HG002	Minimap2	PacBio	Quiver	40.84	99.92	98.21	10	44.55
Human HG002	Minimap2	PacBio	Racon^∗	40.74	99.89	97.34	10	44.30
Human HG002	Minimap2	Illumina	Apollo	40.86	99.92	98.21	10	44.93
Human HG002	BWA-MEM	Illumina	Apollo	40.86	99.92	98.19	10	44.90
Human HG002	BWA-MEM	Illumina	Pilon^∗	40.86	99.92	98.22	10	44.86
Human HG002	Minimap2	PacBio (9 $\times$ )	Apollo	40.62	99.36	83.34	10	37.17
Human HG002	BWA-MEM	PacBio (9 $\times$ )	Apollo	40.62	99.29	82.54	10	36.04
Human HG002	Minimap2	PacBio (9 $\times$ )	Racon	40.95	98.00	78.70	9	33.82
Human HG002	BWA-MEM	PacBio (9 $\times$ )	Racon	40.94	97.27	76.30	9	32.07

Table 17. Table S17: Performance of the aligners

Dataset for	Assembler	Aligner	Platform of the	Number of	Runtime	Memory
the Assembly			Aligned Reads	Alignments		(GB)
E. coli K-12 - ONT	Miniasm	Minimap2	ONT	8,095,856	3m 30s	4.88
E. coli K-12 - ONT	Canu	Minimap2	ONT	1,662,306	39s	2.10
E. coli K-12 - ONT (30 $\times$ )	Canu	Minimap2	ONT (30 $\times$ )	170,910	6s	0.60
E. coli O157 - PacBio	Miniasm	Minimap2	PacBio	732,397	25s	1.79
E. coli O157 - PacBio	Miniasm	Minimap2	Illumina	21,933,051	1m 35s	3.16
E. coli O157 - PacBio	Canu	Minimap2	PacBio	741,343	22s	1.80
E. coli O157 - PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ )	148,241	5s	0.67
E. coli O157 - PacBio (30 $\times$ )	Canu	Minimap2	PacBio (30 $\times$ , Corr)	137,620	3s	0.47
E. coli O157 - PacBio	Miniasm	BWA-MEM	Illumina	19,799,002	2m 34s	3.17
E. coli O157 - PacBio	Canu	BWA-MEM	Illumina	23,328,379	1m 16s	2.89
E. coli O157 - PacBio (30 $\times$ )	Canu	BWA-MEM	Illumina	23,326,202	1m 20s	2.96
E. coli O157 - PacBio	Miniasm	pbalign	PacBio	49,561	12m 55s	6.36
E. coli O157 - PacBio	Canu	pbalign	PacBio	51,994	11m 29s	6.28

Table 18. Table S18: Apollo’s robustness based on the chunk size of the long read and the contig

Long Read	Contig Chunk	Aligned	Aligned	Accuracy
Chunk Size	Size	Bases	Bases (%)
1000	Original	5,708,747	98.49	0.9798
1000	25000	5,487,736	94.46	0.9733
1000	50000	5,689,120	97.95	0.9728
1000	100000	5,493,663	94.52	0.9727
5000	25000	5,430,700	93.06	0.8974
5000	50000	5,411,163	92.68	0.8971
5000	100000	5,516,599	94.49	0.8970
10000	25000	5,415,333	92.65	0.8918
10000	50000	5,423,340	92.75	0.8914
10000	100000	5,474,159	93.61	0.8914

Table 19. Table S19: Apollo’s robustness based on the maximum deletion and filter size parameters

Max	Filter	Aligned	Aligned	Accuracy
Deletion (-d)	Size (-f)	Bases	Bases (%)
3	100	5,699,182	97.91	0.9739
5	100	5,696,138	97.93	0.9735
15	100	5,678,838	97.90	0.9731
3	200	5,705,130	98.12	0.9751
5	200	5,704,582	98.12	0.9750
15	200	5,702,478	98.14	0.9751

Table 20. Table S20: Apollo’s robustness based on the maximum insertion and filter size parameters

Max	Filter	Aligned	Aligned	Accuracy
Insertion (-i)	Size (-f)	Bases	Bases (%)
1	100	5,685,635	97.89	0.9660
5	100	5,638,585	97.62	0.9696
10	100	5,365,978	95.54	0.9531
1	200	5,685,040	98.02	0.9668
5	200	5,692,813	98.07	0.9740
10	200	5,623,736	97.62	0.9701

Table 21. Table S21: Apollo’s robustness based on the match transition, insertion transition probabilities, and the filter size parameters

Match Transition	Insertion Transition	Filter	Aligned	Aligned	Accuracy
Probability (-tm)	Probability (-ti)	Size (-f)	Bases	Bases (%)
0.60	0.25	100	5,670,852	97.95	0.9625
0.60	0.30	100	5,660,957	97.90	0.9596
0.80	0.10	100	5,699,660	98.02	0.9788
0.90	0.05	100	5,685,770	97.89	0.9774
0.60	0.25	200	5,682,512	98.10	0.9644
0.60	0.30	200	5,681,993	98.13	0.9618
0.80	0.10	200	5,707,293	98.16	0.9803
0.90	0.05	200	5,695,902	98.05	0.9789

Table 22. Table S22: List of the parameters that are used to align the reads to the assemblies

Aligner	Parameters
BWA-MEM	-t 45
Minimap2 (for PacBio)	-x map-pb -a -t 45
Minimap2 (for ONT)	-x map-ont -a -t 45
Minimap2 (for Illumina)	-a -x sr -t 45
pbalign	–nproc 45

Equations30

α_{D}^{x} = \frac{f ^{k - x} α _{d e l}}{j = 0 \sum k - 1 f ^{j}} 1 \leq x \leq k

α_{D}^{x} = \frac{f ^{k - x} α _{d e l}}{j = 0 \sum k - 1 f ^{j}} 1 \leq x \leq k

F_{1} (j) = α_{0 j} e_{j} (r [1]) s . t . j \in V_{s}, E_{0 j} \in E_{s}

F_{1} (j) = α_{0 j} e_{j} (r [1]) s . t . j \in V_{s}, E_{0 j} \in E_{s}

F_{t} (j) = i \in V_{s} \sum F_{t - 1} (i) α_{ij} e_{j} (r [t]) j \in V_{s}, 1 < t \leq m

F_{t} (j) = i \in V_{s} \sum F_{t - 1} (i) α_{ij} e_{j} (r [t]) j \in V_{s}, 1 < t \leq m

B_{m} (i) = α_{i (m + 1)} i \in V_{s}, E_{i (m + 1)} \in E_{s}

B_{m} (i) = α_{i (m + 1)} i \in V_{s}, E_{i (m + 1)} \in E_{s}

B_{t} (i) = j \in V_{s} \sum α_{ij} e_{j} (r [t + 1]) B_{t + 1} (j) j \in V_{s}, \leavevmode 1 \leq t < m

B_{t} (i) = j \in V_{s} \sum α_{ij} e_{j} (r [t + 1]) B_{t + 1} (j) j \in V_{s}, \leavevmode 1 \leq t < m

e_{i}^{*} (X) = \frac{t = 1 \sum m F _{t} ( i ) B _{t} ( i ) [ r [ t ] = X ]}{t = 1 \sum m F _{t} ( i ) B _{t} ( i )} \forall X \in {A, C, G, T}, \forall i \in V_{s}

e_{i}^{*} (X) = \frac{t = 1 \sum m F _{t} ( i ) B _{t} ( i ) [ r [ t ] = X ]}{t = 1 \sum m F _{t} ( i ) B _{t} ( i )} \forall X \in {A, C, G, T}, \forall i \in V_{s}

α_{ij}^{*} = \frac{t = 1 \sum m - 1 α _{ij} e _{j} ( r [ t + 1 ]) F _{t} ( i ) B _{t + 1} ( j )}{t = 1 \sum m - 1 x \in V _{s} \sum α _{i x} e _{x} ( r [ t + 1 ]) F _{t} ( i ) B _{t + 1} ( x )} \forall E_{ij} \in E_{s}

α_{ij}^{*} = \frac{t = 1 \sum m - 1 α _{ij} e _{j} ( r [ t + 1 ]) F _{t} ( i ) B _{t + 1} ( j )}{t = 1 \sum m - 1 x \in V _{s} \sum α _{i x} e _{x} ( r [ t + 1 ]) F _{t} ( i ) B _{t + 1} ( x )} \forall E_{ij} \in E_{s}

\overset{e}{^}_{i} (X) = \frac{G _{s} \in S \sum e _{i}^{*, G_{s}} ( X )}{∣ S ∣} \forall X \in Σ

\overset{e}{^}_{i} (X) = \frac{G _{s} \in S \sum e _{i}^{*, G_{s}} ( X )}{∣ S ∣} \forall X \in Σ

\overset{α}{^}_{ij} = \frac{G _{s} \in S \sum α _{ij}^{*, G_{s}}}{∣ S ∣}

\overset{α}{^}_{ij} = \frac{G _{s} \in S \sum α _{ij}^{*, G_{s}}}{∣ S ∣}

v_{1} (j) = \overset{α}{^}_{s t a r t - j} \overset{e}{^}_{j} (X_{j}) \forall j \in V

v_{1} (j) = \overset{α}{^}_{s t a r t - j} \overset{e}{^}_{j} (X_{j}) \forall j \in V

b_{1} (j) = s t a r t \forall j \in V

b_{1} (j) = s t a r t \forall j \in V

v_{t} (j) = i \in V max v_{t - 1} (i) \overset{α}{^}_{ij} \overset{e}{^}_{j} (X_{j}) \forall j \in V, 1 < t \leq T

v_{t} (j) = i \in V max v_{t - 1} (i) \overset{α}{^}_{ij} \overset{e}{^}_{j} (X_{j}) \forall j \in V, 1 < t \leq T

b_{t} (j) = i \in V argmax v_{t - 1} (i) \overset{α}{^}_{ij} \overset{e}{^}_{j} (X_{j}) \forall j \in V, 1 < t \leq T

b_{t} (j) = i \in V argmax v_{t - 1} (i) \overset{α}{^}_{ij} \overset{e}{^}_{j} (X_{j}) \forall j \in V, 1 < t \leq T

v_{T} (e n d) = i \in V max v_{T} (i) \overset{α}{^}_{i - e n d}

v_{T} (e n d) = i \in V max v_{T} (i) \overset{α}{^}_{i - e n d}

b_{T} (e n d) = i \in V argmax v_{T} (i) \overset{α}{^}_{i - e n d}

b_{T} (e n d) = i \in V argmax v_{T} (i) \overset{α}{^}_{i - e n d}

Peer Reviews

No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.

Code & Models

Repositories

CMU-SAFARI/Apollo
noneOfficial

Videos

No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.

Full text

Supplementary Material for

Apollo: A Sequencing-Technology-Independent, Scalable, and Accurate Assembly Polishing Algorithm

Can Firtina, Jeremie S. Kim, Mohammed Alser, Damla Senol Cali, A. Ercument Cicek,

Can Alkan, and Onur Mutlu

1 Constructing a profile hidden Markov model graph

Apollo constructs a profile hidden Markov model graph (pHMM-graph) to represent the sequences of contig as well as the errors that a contig may have. A pHMM-graph includes states and directed transitions from a state to another. There are two types of probabilities that the graph contains: (1) emission and (2) transition probabilities. First, each state has emission probabilities for emitting certain characters where each character is associated with a probability value with the range $[0,1]$ . Each emission probability reveals how likely it is to emit (e.g., consume or output) a certain character when a certain state is visited. Second, each transition is associated with a probability value with the range $[0,1]$ . A transition probability shows the probability of visiting a state from a certain state. Thus, one can calculate the likelihood of emitting all the characters in a given sequence by traversing a certain path in the graph.

The structure of the pHMM-graph allows us to handle insertion, deletion, and substitution errors by following certain states and transitions. Now, we will explain the structure of the graph in detail. For an assembly contig $C$ , let us define the pHMM-graph that represents the contig $C$ as $G(V,E)$ . Let us also define the length of the contig $C$ as $n=|C|$ . A base $C[t]$ has one of the letters in the alphabet set $\Sigma=\{A,C,G,T\}$ . Thus, a state emits one of the characters in $\Sigma$ with a certain probability. For a state $i$ , We denote the emission probability of a base $c\in\Sigma$ as $e_{i}(c)\in[0,1]$ where $\sum\limits_{c\in\Sigma}e_{i}(c)=1$ . We denote the transition probability from a state, i, to another state, j, as $\alpha_{ij}\in[0,1]$ . For the set of the states that the state $i$ has an outgoing transition to, $V_{i}$ , we have $\sum\limits_{j\in V_{i}}\alpha_{ij}=1$ . Now let us define in four steps how Apollo constructs the states and the transitions of the graph $G(V,E)$ :

First, Apollo constructs a start state, $v_{start}\in V$ , and an end state $v_{end}\in V$ . Second, for each base $C[t]$ where $1\leq t\leq n$ , Apollo constructs a match state as follows (Figure S1):

•

A match state that we denote as $M_{t}$ for the base $C[t]$ where $M=C[t]$ s.t. $C[t]\in\Sigma$ and $M_{t}\in V$ (i.e., if the $t^{th}$ base of the contig $C$ is $G$ , then the corresponding match state is $G_{t}$ ). For the following steps, let us assume $i=M_{t}$

•

A match emission with the probability $\beta$ , for the base $C[t]$ s.t. $e_{i}(C[t])=\beta$ . $\beta$ is a parameter to Apollo.

•

A substitution emission with the probability $\delta$ , for each base $c\in\Sigma$ and $c\neq C[t]$ s.t. $e_{i}(c)=\delta$ (Note that $\beta+3\delta=1$ ). $\delta$ is a parameter to Apollo.

•

A match transition with the probability $\alpha_{M}$ , from the match state $M_{t}=i$ to the next match state $M_{t+1}=j$ s.t. $\alpha_{ij}=\alpha_{M}$ . $\alpha_{M}$ is a parameter to Apollo.

Third, for each base $C[t]$ where $1\leq t\leq n$ , Apollo constructs the insertion states as follows (Figure S2):

•

There are $l$ many insertion states, $I_{t}^{1}$ , $I_{t}^{2}$ , …, $I_{t}^{l}$ , where $I_{t}^{i}\in V$ , $1\leq i\leq l$ and $l$ is a parameter to Apollo

•

The match state, $M_{t}=i$ , has an insertion transition to $I_{t}^{1}=j$ , with the probability $\alpha_{I}$ s.t. $\alpha_{ij}=\alpha_{I}$

•

For each $i$ where $1\leq i<l$ , the insertion state $I_{t}^{i}=k$ has an insertion transition to the next insertion state $I_{t}^{i+1}=j$ with the probability $\alpha_{I}$ s.t. $\alpha_{kj}=\alpha_{I}$

•

For each $i$ where $1\leq i<l$ , the insertion state $I_{t}^{i}=k$ has a match transition to the match state of the next base $M_{t+1}=j$ with the probability $\alpha_{M}$ s.t. $\alpha_{kj}=\alpha_{M}$

•

The last insertion state, $I_{t}^{l}$ , has no further insertion transitions. Instead, it has a transition to the match state of the next base $M_{t+1}=j$ with the probability $\alpha_{M}+\alpha_{I}$ s.t. $\alpha_{kj}=\alpha_{M}+\alpha_{I}$

•

For each $i$ where $1\leq i\leq l$ , each base $c\in\Sigma$ and $c\neq C[t+1]$ has an insertion emission probability $1/3\approx 0.33$ for the insertion state $I_{t}^{i}=k$ s.t. $e_{k}(c)=0.33$ and $e_{k}(C[t+1])=0$ . Note that $\sum\limits_{c\in\Sigma}e_{k}(c)=1$ . (i.e., if the base at the location $t+1$ is T, then $e_{k}(A)=0.33$ , $e_{k}(T)=0$ , $e_{k}(G)=0.33$ , and $e_{k}(C)=0.33$ ).

Fourth step for finalizing the complete structure of the pHMM graph, for each state $i\in V$ , Apollo constructs the deletion transitions as follows (Figure S3):

•

Let us define $\alpha_{del}=1-(\alpha_{M}-\alpha_{I})$ , which is the overall deletion transition probability.

•

There are $k$ many deletion transitions from the state $i$ , to the further match states. $k$ is a parameter to Apollo.

•

We assume that a transition deletes the bases if it skips the corresponding match states of the bases. We denote the transition probability of a deletion transition as $\alpha_{D}^{x}$ s.t. $1\leq x\leq k$ , if it deletes $x$ many bases in a row in one transition. Apollo calculates the deletion transition probability $\alpha_{D}^{x}$ using the normalized version of a polynomial distribution where $f\in[0,\infty)$ is a factor value for the equation:

[TABLE]

•

If the $f$ value is set to $1$ , then the each deletion transition is equally likely (i.e., $\alpha_{D}^{1}=\alpha_{D}^{10}$ , if $k\geq 10$ ). As the $f$ value increases, the probability of deleting more bases in one transition decreases accordingly (i.e., $\alpha_{D}^{1}\gg\alpha_{D}^{10}$ , if $k\geq 10$ ). $f$ is a parameter to Apollo.

We note that the start state $v_{start}$ also has a match transition to $M_{1}$ and deletion transitions as defined previously. There are al $l$ many insertion states, $I_{0}^{1}$ , $I_{0}^{2}$ , …, $I_{0}^{l}$ , between the start state and the first match state $M_{1}$ . The transitions of these insertion states are also identical to what we described before. We would also like to note that the end state $v_{end}$ has no outgoing transition. The prior states consider $v_{end}$ as a match state and connect to it accordingly. The start and end states have no emission probabilities.

Note that the design of pHMM-graph described here and proposed in Hercules [1] is different from the conventional pHMM-graphs [2]. One significant difference is that the conventional pHMM-graphs have deletion states for each match state whereas the pHMM-graph model of Apollo uses deletion transitions instead of states. In the conventional model, visiting deletion states does not consume (i.e., emit) a character from a given sequence (i.e., observation). Therefore, this requires storing extra "position" information that tells which character should be consumed given a state at iteration $i$ (i.e., in each transition from a state to another). We want to make sure that each state consumes only one character (and no more) when visited to prevent storing the extra position information. In Apollo’s design, iteration number $i$ equals the position of a character that is being consumed Apollo’s states consume exactly one character. This allows us to remove an entire dimension, the iteration number $i$ , which greatly helps us to reduce both memory requirements and runtime while calculating the Forward-Backward values.

2 The Forward-Backward and Baum-Welch Algorithms

Apollo uses the region of a pHMM-graph (i.e., sub-graph) that a read (i.e., observation or a sequence) is aligned to in order to calculate the likelihood of each state emitting a certain base at position $t$ in the aligned read. However, this does not mean that position $t$ is known since we need to consider the fact that an unknown number of insertion and deletion errors may have occurred when $k$ number of transitions is followed from the start state to a certain state. Therefore, states should be measuring the likelihood of emitting a character at position $t$ where $t$ is a number in range $[1...k]$ where $k$ is the number of transitions that was taken so far. In the no error case, we have $k=t$ . Apollo uses reads as observations for the Forward-Backward algorithm [3] in order to calculate the likelihoods per state. These likelihoods are calculated based on initial transition and emission probabilities of a pHMM-graph and the read itself. Apollo uses these likelihoods to make the contig similar to the aligned read. Apollo, then, trains the pHMM-graph of a contig per each read that aligns to the contig using the Baum-Welch algorithm [3]. We describe the details of both the Forward-Backward and the Baum-Welch algorithms in the following paragraphs.

For each read aligning to a contig, Apollo uses the alignment location and the sequence of the read in order to train the pHMM-graph. First, per each aligned read sequence $r$ , Apollo extracts the sub-graph $G_{s}(V_{s},E_{s})$ that corresponds to the aligned region of the contig where we have $v_{start}$ , $v_{end}$ , match and insertion states, and the transitions as described in the Supplementary Section 1. Each transition from state $i\in V_{s}$ to state $j\in V_{s}$ , $E_{ij}\in E_{s}$ , is associated with a transition probability $\alpha_{ij}$ . For every pair of states, $i\in V_{s}$ and $j\in V_{s}$ , the transition probability $\alpha_{ij}=0$ if $E_{ij}\not\in E_{s}$ . Let us define the length of the aligned read, $r$ , as $m=|r|$ . Second, it calculates the forward and backward probabilities of each state based on the aligned read, $r$ .

Let us assume that the forward probability of a state $j$ that observes the $t^{th}$ base of the aligned read, $r[t]$ , is $F_{t}(j)$ . For the forward probability, observing the $t^{th}$ base at the state $j$ means that all the previous bases ( $r[1]\dots r[t-1]$ and $1<t\leq m$ ) have been observed by following a path starting from the start state to the state $j$ and $j$ observes the next base, $r[t]$ . All possible transitions that lead to state $j$ to observe the base $r[t]$ contribute to the probability with (1) the forward probability of the origin state $i$ calculated with the $(t-1)^{th}$ base of $r$ , $F_{t-1}(i)$ , (2) multiplied by the probability of the transition from $i$ to $j$ , $\alpha_{ij}$ , (3) multiplied by the probability of emitting the base $r[t]$ at state $j$ , $e_{j}(r[t])$ .

Let us denote the start state $v_{start}$ with the index value of [math] (i.e., $v_{start}=0$ ). For each state $j\in V_{s}$ , we calculate the forward probability, $F_{t}(j)$ , as follows where $F_{1}(j)$ is the initialization step:

[TABLE]

Let us assume that the backward probability of a state $i$ that observes $t^{th}$ base of the aligned read, $r[t]$ , is $B_{t}(i)$ . For the backward probability, observing the $t^{th}$ base at the state $i$ means that all the further bases ( $r[t+1]\dots r[m]$ and $1\leq t<m$ ) have been observed by following a path starting from the end state to the state $i$ (backwards) and $i$ observes the previous base, $r[t]$ . All possible transitions that lead to state $i$ to observe the base $r[t]$ contribute to the probability with (1) the backward probability of the next state $j$ calculated with the $(t+1)^{th}$ base of $r$ , $B_{t+1}(j)$ , (2) multiplied by the probability of the transition from $i$ to $j$ , $\alpha_{ij}$ , (3) multiplied by the probability of emitting the base $r[t+1]$ at state $j$ , $e_{j}(r[t+1])$ .

Let us denote the end state $v_{end}$ with the index value of $m+1$ (i.e., $v_{end}=m+1$ ). For each state $j\in V_{s}$ , we calculate the backward probability, $B_{t}(i)$ , as follows where $B_{m}(i)$ is the initialization step:

[TABLE]

The calculations of forward and backward probabilities are referred as the Forward-Backward algorithm. After calculation of the forward and backward probabilities, Apollo uses the Baum-Welch algorithm to train the pHMM-graph by calculating the posterior transition and the emission probabilities of the sub-graph, $G_{s}$ , as shown in equations S4 and S5, respectively. In equation S4, we use the Iversonian brackets [4] to denote that $[r[t]=X]$ is $1$ if the $t^{th}$ character of $r$ is the same character as $X$ . Otherwise, $[r[t]=X]$ is [math]. This structure helps us to perform the summation in the numerator only when the character at a position equals to the character given in function $e^{*}_{i}(X)$ (i.e., $X$ ). We, then, normalize this summation to make sure the sum of the emission probabilities that state $i$ can have is equal to 1.

[TABLE]

3 Joining Posterior Probabilities

As we explain in the Supplementary Section 2, for each read that aligns to the contig, Apollo extracts a sub-graph $G_{s}$ and uses the Forward-Backward algorithm to train the sub-graph. It is highly possible that there can be overlaps between two or many sub-graphs such that the sub-graphs can include the same states and the transitions when using high coverage reads. However, the updates on the overlapping states and the transitions are exclusive between the sub-graphs such that no two update in separate graphs affect each other while calculating the Forward or the Backward probabilities. Each sub-graph uses the initial probabilities to calculate the posterior probabilities. In order to handle training of the overlapping states and the transitions, Apollo takes the average of the posterior probabilities and reports the average probability as the final posterior probability for the entire pHMM-graph.

Let us assume that the set of sub-graphs $S$ includes the same state $i\in V$ . For each $G_{s}$ in $S$ , we obtain a $e^{*}_{i}(X)$ , where $\forall X\in\Sigma$ , which denotes the posterior emission probability as we explain in the Supplementary Section 2. We denote $e^{*}_{i}(X)$ that belongs to $G_{s}$ as $e^{*,G_{s}}_{i}(X)$ . Then, Apollo finds the final emission value $\hat{e}_{i}(X)$ as follows:

[TABLE]

Similarly, let us assume that the set of sub-graphs $S$ includes the same transition edge $E_{ij}\in E$ . For each $G_{s}$ in $S$ , we obtain an $\alpha^{*}_{ij}$ that denotes the posterior transition value. We define $\alpha^{*}_{ij}$ that belongs to $G_{s}$ as $\alpha^{*,G_{s}}_{ij}$ . Apollo finds the final transition value $\hat{\alpha}_{ij}$ as follows:

[TABLE]

If a state in $V$ or an edge in $E$ is not covered by a read then Apollo retains the initial emission and transition probabilities and uses as posterior probabilities, respectively.

We would like to note that the Baum-Welch algorithm is also used to train conventional hidden Markov models (HMMs). In each observation, the Baum-Welch algorithm updates the transition and emission probabilities of an HMM accordingly. The initial probabilities of such HMMs may even be assigned randomly. This means that the order of the observations (i.e., training data), and the initial probabilities used to train an HMM also affect the overall accuracy as the following observations usually use the HMM that is trained based on earlier observations. Therefore, after using all the training data, an HMM may still have room to converge to a local optimal point due to the biases caused by the initial probabilities and the order of the training data. The usual approach to mitigate such biases is to train HMMs multiple times until the overall accuracy of an HMM converges to a certain point. We do not follow this strategy because of three reasons. First, Apollo does not set the initial transition and emission probabilities randomly. Instead, the probabilities are usually set according to the error profile of an assembly. Second, we use the initial probabilities each time a read is used to train the pHMM-graph so the order of the training data does not matter. Third, Apollo is a very time consuming tool and taking multiple iterations until convergence would significantly increase the overall runtime, which we want to avoid.

4 Decoding with the Viterbi Algorithm

Apollo uses the Viterbi algorithm [5] to reveal the polished assembly by finding the most likely path starting from the start state, $v_{start}$ , of the trained graph $G$ to the end state, $v_{end}$ . For each state $j$ , the Viterbi algorithm calculates $v_{t}(j)$ , which is the maximum marginal forward probability $j$ obtained from following a path starting from the start state when decoding the $t^{th}$ base of the polished contig. Let $X_{j}\in\Sigma$ be the base that has the greatest emission probability for the state $j$ , i.e., $\hat{e}_{j}(X_{j})\geq\hat{e}_{j}(x)$ , $\forall x\in\Sigma$ . Then, the value of $v_{t}(j)$ depends on 1) the transition probability from state $i$ to the state $j$ , $\hat{\alpha}_{ij}$ , 2) the Viterbi value of the state $i$ when decoding the $(t-1)^{th}$ base of the polished contig, $v_{t-1}(i)$ , and 3) the emission probability of the base $X_{j}$ , $\hat{e}_{j}(X_{j})$ . The Viterbi algorithm also keeps a back pointer, $b_{t}(j)$ , which keeps track of the predecessor state $i$ that yields the $v_{t}(j)$ value.

Let $T$ be the length of the decoded sequence, which is initially unknown. The algorithm recursively calculates $v$ values for each position $t$ of a decoded sequence as described in the equations S8.1 and S8.3. The algorithm stops at iteration $T^{*}$ such that for the last $iter$ iterations, the maximum value we have observed for $v(end)$ cannot be improved and $iter$ is set to 50 by default (empirically chosen). $T$ is then set to $t^{*}$ such that $v_{t^{*}}(end)$ is the maximum among all iterations $1\leq t\leq T^{*}$ .

Initialization

[TABLE]

[TABLE] 2. 2.

Recursion

[TABLE]

[TABLE] 3. 3.

Termination

[TABLE]

The polished contig is generated by recursively following states from the end state, $v_{end}$ , at time $T$ until the back pointer points back to the start state, $v_{start}$ , at time $t=1$ for the state $j$ as follows:

5 Performance of the Assembly Polishing Algorithms

In Tables S1, S2, S3, S6 S9, and S12, we compare the assembly polishing performance of Apollo to the competing algorithms based on the difference between the assemblies and their reference genomes (i.e., ground truth). In Tables S4, S7, S10, S13, and S15, we show the k-mer similarities between Illumina reads and the assemblies to provide an alignment-free comparison between the tools. We also use QUAST [6] to make a more detailed quality assessment of the assemblies in Tables S5, S8, S11, S14, and S16.

6 Performance of the Aligners

Here in Table S17, we show the performances of the aligners in terms of number of alignments that the aligners generate given the assembly and the reads to align, runtime (wall clock), and the memory requirement.

7 Robustness of Apollo

Here in Tables S18, S19, S20, S21, we show the robustness of Apollo based on the parameters that has a direct affect on the machine learning algorithm. In each of the tables we show that Apollo is robust to different set of parameters.

8 Parameters

We show the parameter settings of the aligners that we used to align the reads to the assembly in Table S22.

Bibliography10

The reference list from the paper itself. Each links out to its DOI / PubMed record.

1[1] Can Firtina, Ziv Bar-Joseph, Can Alkan, and A Ercument Cicek. Hercules: a profile HMM-based hybrid error correction algorithm for long reads. Nucleic Acids Research , 46(21):e 125–e 125, August 2018.
2[2] Sean R. Eddy. Profile hidden Markov models. Bioinformatics , 14(9):755–763, October 1998.
3[3] L. E. Baum. An inequality and associated maximization technique in statistical estimation of probabilistic functions of a Markov process. Inequalities , 3:1–8, 1972.
4[4] Donald E. Knuth. Two Notes on Notation. The American Mathematical Monthly , 99(5):403, May 1992.
5[5] A. Viterbi. Error bounds for convolutional codes and an asymptotically optimum decoding algorithm. IEEE Transactions on Information Theory , 13(2):260–269, April 1967.
6[6] Alexey Gurevich, Vladislav Saveliev, Nikolay Vyahhi, and Glenn Tesler. QUAST: quality assessment tool for genome assemblies. Bioinformatics , 29(8):1072–1075, April 2013.
7[7] Sergey Koren, Brian P. Walenz, Konstantin Berlin, Jason R. Miller, Nicholas H. Bergman, and Adam M. Phillippy. Canu: scalable and accurate long-read assembly via adaptive k -mer weighting and repeat separation. Genome Research , 27(5):722–736, May 2017.
8[8] Heng Li. Minimap and miniasm: fast mapping and de novo assembly for noisy long sequences. Bioinformatics , 32(14):2103–2110, July 2016.