Evaluation of the effectiveness of semen storage and sperm purification methods for spermatozoa transcript profiling

Figures & data

Figure 1.  Flow diagram of sperm sample processing. Semen samples were classified as liquefied storage (L) or pelleted storage (P); for each storage, the samples were divided into two aliquots after initial cell counting and purified by either PureSperm (PS) or somatic cell lysis buffer (SCLB) methods. Final cell count determined the cell loss after purification; equal amount of cells were used as cell input for RNA isolation. Next, equal amount isolated RNA from each sample was applied as input for RNA-Seq library preparation; the libraries were deep sequenced for 50 cycles paired-end; the sequencing reads were demultiplexed, and short reads were aligned to human reference genome. Finally the alignment results were analyzed. The effect of sample storage and sperm purification methods to RNA profiling was evaluated.

Table 1. Cell recovery and yield of RNA from PS and SCLB purification methods.

	PureSperm (PS)				Somatic Cell Lysis Buffer (SCLB)
Samples	Input cells	Recovered cells	Recovery rate	RNA (ng) / 10^6 cells	Input cells	Recovered cells	Recovery rate	RNA (ng) / 10^6 cells
S1^P	166.5	78	46.8%	37	166.5	132	79.3%	14
S2^P	145	96	66.2%	47	145	108	74.5%	12
S3^P	132.5	78	58.9%	86	132.5	120	90.7%	14
S4^P	155	66	42.6%	23	155	132	85.2%	19
S1^L	150	98	65.3%	29	150	120	80.0%	20
S2^L	80	56	70.0%	26	80	94*	100.0%	38
S3^L	150	84	56.0%	89	150	132	88.0%	30
S4^L	70	70	100.0%	76	70	70	100.0%	15

The values in columns 'input cells‘ and ’recovered cells‘ are as million (x 10⁶). The samples from four subjects were denoted S₁, S₂, S₃, and S₄. P: pelleted storage; L: liquefied storage; Input cells: the number of sperm cells as initial input for purification; Recovered cells: the number of recovered sperm after purification; Recovery rate: the percentage of recovery; RNA (ng) / 10⁶ cells: the yield of isolated RNA (ng) from one million sperm cells; *: in one SCLB sample, the number of recovered cells is higher than the number of input cells. Probably this is related with sperm cell counting method. Its recovery rate is reset back to 100% in the table.

Table S1. The number of sperm from each sample (in million cells)

Subjects	Pellet storage (P)	Liquefied storage (L)
S1	333	380
S2	290	160
S3	265	510
S4	310	140

Figure 2.  Distribution of RNA-seq fragment sizes of 16 samples. Fragment sizes for each paired-end sequencing library were inferred from the separation of read pairs, which were mapped to the human reference genome using novoalign. It is clear that the average RNA fragments from samples purified by Pure Sperm (Blue and Green) is longer than that from SCLB (Red and Orange). P: pelleted storage; L: liquefied storage; PS: PureSperm purification; SCLB: somatic cell lysis buffer purification.

Table S2. RNA-seq statistics summary

Samples	Total reads	Aligned reads	% of align	Unique aligned reads	Average length (bp)	STDEV (bp)
S1^P-SCLB	16.9	14.2	83.8%	10.0	85	35.2
S2^P-SCLB	14.5	13.2	90.9%	11.2	93	40.3
S3^P-SCLB	14.7	13.4	90.8%	11.9	86	33.6
S4^P-SCLB	17.4	14.8	85.4%	12.0	90	39.4
S1^L-SCLB	17.9	14.4	80.5%	11.5	92	37.6
S2^L-SCLB	21.2	19.3	91.1%	15.5	93	38.8
S3^L-SCLB	42.0	37.2	88.5%	31.3	93	39.1
S4^L-SCLB	19.1	16.9	88.7%	14.0	96	43.4
S1^P-PS	29.2	27.1	92.8%	22.6	110	37.9
S2^P-PS	32.4	29.6	91.3%	26.1	108	41.1
S3^P-PS	26.7	25.0	93.7%	22.3	111	39.7
S4^P-PS	24.9	22.9	91.7%	19.8	106	40.5
S1^L-PS	35.9	34.8	97.0%	28.8	103	38.2
S2^L-PS	27.8	26.4	95.1%	23.3	106	39.9
S3^L-PS	25.8	23.7	92.1%	20.8	105	39.9
S4^L-PS	27.0	26.0	96.3%	22.5	105	38.7

The read counts in the table are in millions of reads. “Total reads” indicates the number of reads generated from sequencer; “Aligned reads” indicate the number of reads that have been aligned to reference genome by aligner; “Unique aligned reads” indicates the number of reads that were uniquely aligned to reference genome. “Average length” and “STDEV” are the average length of mapped fragments and their standard deviation. S1∼S4 are four different subjects. P-PS denotes pelleted storage with PureSperm purification; L-PS denotes liquefied storage with PureSperm purification; P-SCLB denotes pelleted storage and SCLB purification; L-SCLB denotes liquefied storage and SCLB purification.

Table 2. Reads mapped to mitochondria genome and 12S, 16S rRNA.

Samples	Unique aligned reads	Mapped to mitochondria	% of Mito	Mapped to 12S	% of 12S	Mapped to 16S	% of 16S
S1^P-SCLB	10.0	0.18	1.79%	0.11	59.2%	0.062	34.7%
S2^P-SCLB	11.2	0.12	1.06%	0.077	65.2%	0.036	30.7%
S3^P-SCLB	11.9	0.45	3.80%	0.36	78.5%	0.09	19.9%
S4^P-SCLB	12.0	0.045	0.37%	0.022	50.3%	0.02	44.5%
S1^L-SCLB	11.5	0.16	1.42%	0.091	56.0%	0.06	38.3%
S2^L-SCLB	15.5	0.2	1.30%	0.12	57.9%	0.08	39.5%
S3^L-SCLB	31.3	2.0	6.27%	1.5	74.4%	0.47	24.1%
S4^L-SCLB	14.0	0.05	0.38%	0.03	56.9%	0.017	32.3%
S1^P-PS	22.6	13.7	60.5%	8.0	58.7%	5.5	40.2%
S2^P-PS	26.1	14.3	54.8%	9.8	68.7%	4.4	30.6%
S3^P-PS	22.3	12.2	55.0%	9.4	77.2%	2.7	21.9%
S4^P-PS	19.8	8.9	45.0%	5.5	61.2%	3.3	37.2%
S1^L-PS	28.8	19.4	67.2%	11.5	59.2%	7.7	39.8%
S2^L-PS	23.3	14.7	63.0%	8.8	59.9%	5.7	38.6%
S3^L-PS	20.8	12.8	61.2%	9.7	75.9%	3.0	23.4%
S4^L-PS	22.5	14.8	65.7%	10.1	68.3%	4.5	30.7%

The read counts in the table are in millions of reads. Unique aligned reads indicates the number of reads that are uniquely aligned to human reference genome; Mapped to mitochondria indicates the number of uniquely aligned reads that mapped to mitochondria; % of Mito indicates that percentage of reads that mapped to mitochondria; Mapped to 12S and Mapped to 16S indicate the number of reads mapped to 12S and 16S mitochondrial ribosomal RNA. S1∼S4 are the samples from four subjects; P: pelleted storage; L: liquefied storage; PS: PureSperm purification; SCLB: somatic cell lysis buffer purification

Figure 3.  Unsupervised clustering of the four storage / purifcation methods. In each sample, the transcript profile generated from four methods were clustered. The Y-axis is the 1 - Pearson correlation coefficient. S1 ∼ S4: samples from four subjects; P: pelleted storage; L: liquefied storage; PS: PureSperm purification; SCLB: somatic cell lysis buffer purification.

Figure 4.  Number of stable transcript pairs as a function of different criteria combination. The Y-axis is the number of stable transcript pairs obtained. In Panel A, the X-axis is Pearson correlation coefficient (Pcc); in Panel B, the X-axis is Coefficient of Variation (CofV); in Panel C, the X-axis is FPKM (fragments per kilobase of exon per million fragments mapped) values. A stable pair also needs to satisfy the following criteria: In panel A, CofV ≤ 0.2, FPKM ≥ 1; in panel B, Pcc ≥ 0.9, FPKM ≥ 1; and in panel C, Pcc ≥ 0.9, CofV ≤ 0.2. The Spearman correlation coefficient = 1 for the stable pairs in all three panels.

Figure 5.  The distribution of reads mapping to gene RPS4X, which was generated from UCSC genome browser by uploading the corresponding short read positions. The blue arrow indicates the transcript 5' to 3' orientation. In Gene RPS4X, the thicker rectangles indicate the exons, the thinner rectangles at two ends indicate 5' UTR and 3' UTR. The number of reads corresponding to each base position is represented on the vertical axis. S1 ∼ S4: samples from four subjects; P: pelleted storage; L: liquefied storage; PS: PureSperm purification; SCLB: somatic cell lysis buffer purification

Figure S1.  PCR assessment of DNA contamination and RNA integrity. A. Isolated RNAs and two human genomic cDNA samples were used as template. B. The synthesized cDNAs were used as template. The primer pair spans the PRM1 intron. S1∼S4 are the samples from four subjects; ‘P’ denotes pelleted storage. ‘L’ denotes liquefied storage; PS denotes PureSperm purification and SCLB is somatic cell lysis buffer purification; HG denotes human genome DNA; Ctrl denotes negative control.

Figure S2.  Distribution of reads mapped to mitochondrial genome. The scale for each sample is based on the maximum read count with the displayed region. In the mitochondrial genome, the coordinate of 12S rRNA is 648 – 1,601 and 16S rRNA is 1,671 – 3,229. S1∼S4 are the samples from four subjects; ‘P’ denotes pelleted storage. ‘L’ denotes liquefied storage; PS denotes PureSperm purification and SCLB is somatic cell lysis buffer purification.

Figure S3.  Unsupervised clustering by subjects. In each method, the transcript profile generated from four subjects were clustered. The Y-axis is the 1 - Pearson correlation coefficient. Panel A indicates SCLB purification from liquefied storage (L) samples; Panel B indicates SCLB purification from pelleted storage (P) samples; Panel C indicates PureSperm purification from liquefied storage (L) samples; and Panel D indicate PureSperm purification from pelleted storage (P) samples.

Figure S4.  Number of stable transcript pairs as a function of different criteria. The short reads uniquely aligned to reference genome without mitochondrial sequence were used as input to calculate each gene's FPKM values. The Y-axis is the number of stable transcript pairs. The X-axis is: Pcc in panel A, Cofv in panel B; and FPKM in panel C. The other criteria are also required for stable transcripts identification: In panel A, CofV≤0.2, FPKM≥1; in panel B, Pcc≥0.9, FPKM≥1; in panel C, Pcc≥0.9, CofV≤0.2. The Spearman correlation coefficient=1 for the stable pairs in all three panels.

Table 3. Preferentially isolated genes in PS groups over SCLB groups.

Gene symbol	Gene name	Log2(FC)
RPS4X	ribosomal protein S4	8.44
EHF	ets homologous factor	8.24
SEMG1	semenogelin I	7.78
LOC653881	60S ribosomal protein L3-like	7.34
MTRNR2L6	MT-RNR2-like 6	7.30
RPL13AP5	ribosomal protein L13a pseudogene 5	7.20
NKX3-1	NK3 homeobox 1	7.03
LOC100505479	hypothetical LOC100505479	6.79
MTRNR2L10	MT-RNR2-like 10	6.65
LOC100289130	hypothetical LOC100289130	6.64
MTRNR2L9	MT-RNR2-like 9	6.58
MTRNR2L2	MT-RNR2-like 2	6.57
MTRNR2L8	MT-RNR2-like 8	6.38
RPS10	ribosomal protein S10	6.35
RPL11	ribosomal protein L11	6.15
BASP1	brain abundant, membrane attached signal protein 1	6.01
LOC100287803	hypothetical LOC100287803	5.62
RAB7A	RAB7A, member RAS oncogene family	5.39
MTRNR2L1	MT-RNR2-like 1	4.91

The gene symbols and gene names of 19 preferentially isolated genes in PureSperm (PS) samples over somatic cell lysis buffer (SCLB) samples were listed. Column Log₂(FC) refers to the average of log₂ based fold difference between PureSperm as compared to SCLB. For any preferentially isolated gene, if its log₂(FC) value is infinity in any sample, its value was assigned a value of 10 prior to average.

Table 4. FPKM values of markers in each sperm sample.

Samples	SEMG1	SEMG2	MSMB	PIP	CD34	CD45	CDH1	KIT
S1^P-SCLB	0.0	10.1	0.0	0.0	0.0	0.4	0.1	3.5
S2^P-SCLB	0.3	5.8	0.2	0.0	0.0	4.1	0.0	3.6
S3^P-SCLB	0.0	7.3	2.7	0.0	0.8	2.8	0.2	1.1
S4^P-SCLB	0.3	5.7	0.6	0.0	0.0	0.1	0.0	1.5
S1^L-SCLB	1.4	12.4	0.0	0.0	0.0	27.3	0.0	5.5
S2^L-SCLB	0.4	10.0	3.7	0.0	0.0	0.5	0.0	4.1
S3^L-SCLB	0.0	6.8	0.0	0.0	0.0	0.0	0.0	1.7
S4^L-SCLB	0.0	14.7	0.0	0.0	0.1	6.0	0.7	3.5
S1^P-PS	15.9	14.1	23.6	0.0	0.0	9.6	0.0	0.2
S2^P-PS	14.0	16.8	8.9	0.0	0.3	0.2	0.0	2.0
S3^P-PS	9.8	6.5	0.1	0.0	1.1	2.2	0.8	0.8
S4^P-PS	58.9	34.2	3.7	0.9	0.0	0.4	2.0	9.2
S1^L-PS	55.4	59	67.6	24.3	0.6	1	8.3	9.8
S2^L-PS	32	15.5	17.1	1.2	0.2	4.1	0	1.9
S3^L-PS	11.4	7.8	1.2	0	1.4	2.3	0.9	1
S4^L-PS	35	23.3	1.3	0	0	7.9	0.1	0.6

The values in the table are FPKM values for each marker gene in every sample. S₁∼S₄ are four different subjects. P-PS: pelleted storage and PureSperm purification; L-PS: liquefied storage and PureSperm purification; P-SCLB: pelleted storage and SCLB purification; L-SCLB: liquefied storage and SCLB purification. SEMG1, SEMG2, MSMB, PIP: transcripts encoding proteins secreted by the epididymis and the prostate/seminal vesicles; CD34, CD45: white blood cell marker transcripts; CDH1: epithelial cells transcript; KIT: immature germ cell transcript.