ion torrent proton short reads contamination

[zcm2403]

i had ran a PI chip yesterday.And today i found that there was something wrong in the report.Just as shown in the pdf file,a short reads-about 30 bps was mixed into the final library.could anybody give me some suggestion about the origin of the short reads.Thanks!

[r.rosati]

It might be spurious amplification... have you considered extracting those short-length reads from the BAM file, i.e. with a Python script, and see where they map?

[r.rosati]

...for example, these are (humble) Python 3 scripts that use pysam, pandas and plotly:

first I would check the exact position of the short reads peak:

import pysam
import pandas as pd
from plotly.offline import plot
import plotly.graph_objs as go

bamfiles = ['IonXpress_003.bam',
            'IonXpress_004.bam',
            'IonXpress_007.bam',
            'IonXpress_005.bam'
            ]


# length_data will hold data counts for every bam file
length_data = []

# collect sequence length counts
for file in bamfiles:
    print("Reading file: " + file)
    bam = pysam.AlignmentFile(file, "rb")
    selected = bam.fetch()
    # alen = length of the aligned reference segment
    # qlen = length of the aligned read segment
    # rlen = length of the aligned read segment, including soft-clipped bases
    reads_xlen = [read.alen for read in selected]
    counts = [(i, reads_xlen.count(i)) for i in range(min(reads_xlen), max(reads_xlen))]
    length_data.append(pd.DataFrame(counts))

# create traces
data = [go.Scatter(x=bamdata[0], y=bamdata[1], mode = 'lines', name=bamfiles[i]) for i, bamdata in enumerate(length_data)]
    
# plot data
scatterplot = plot({'data': data})

See exactly at which nucleotide the peak is, and then check if there is any similarity within the most common sequences...

import pysam

# choose a file that contains the peak, select the peak position
file = 'IonXpress_004.bam'
peakpos = 29 # let's say the peak was at 29 bases

# retrieve all sequences from reads with the alignment length you specified
bam = pysam.AlignmentFile(file, "rb")
selected = bam.fetch()
reads = [read for read in selected if read.alen == peakpos]
reads_seq = [read.seq for read in reads]

# you can count them
print("Number of sequences retrieved:", len(reads_seq))

# see how many different entries you have
entries = set(reads_seq)
print("Number of different entries:", len(entries))

# categorize them
cat_entries = [(entry, reads_seq.count(entry)) for entry in entries]
cat_entries.sort(key = lambda x:x[1])

# print the 20 most common sequences
print('\nMost common sequences:')
print('[hits] sequence')
for i in range(20):
    print('[{0}] {1}'.format(cat_entries[i][1], cat_entries[i][0]))

[zcm2403]

Thanks for your help!I had send the bam files to the technical support.His anwser was that the short reads were primer dimers.He suggested us to optimize the experiment protocal.So we will improve the experiment pipeline and run one sample on the PGM to test the propose!