Biology subjects

Rasmussen, M. H.

Publications and source records attributed to Rasmussen, M. H..

9 recordsLinked to original sources

bioRxiv2026

SNV and indel error modeling of deep targeted cell-free DNA sequencing data for sensitive detection of circulating tumor DNA in colorectal cancer

Circulating tumor DNA (ctDNA) is a promising biomarker for cancer detection, but low tumor burden makes it difficult to distinguish true signal from background noise. To aggregate and better evaluate weak mutational signals, we propose PyDREAMS, which incorporates both single-nucleotide variants (SNVs) and insertion and deletions (indels) for ctDNA detection and quantification. To distinguish signal from noise, a neural network background error model is learned from healthy controls. It captures the joint effects of cell-free DNA (cfDNA)-specific lesions and sequencing errors, accounting for both genomic context and read-level features. Finally, a statistical test is used to evaluate the presence of mutational signals. We evaluate the method in a tumor-informed setting, using cohorts of colorectal cancer samples with deep targeted plasma cfDNA sequencing across 12 cancer driver genes. We trained PyDREAMS on 46 healthy controls, with feature analysis revealing that both SNV and indel error rates were lowest at mononucleosomal fragment lengths, suggesting that nucleosomes protect cfDNA and reduce lesion accumulation during circulation and sample handling. In the validation cohort, combining SNVs with indels improved detection, with indels contributing approximately 1.5-fold more evidence per mutation than SNVs. On a test cohort of 209 stage I-III colorectal cancer (CRC) patients and 24 healthy controls, PyDREAMS outperformed a Shearwater-based caller (AUC 0.917 vs 0.909). In stage III post-operative (post-OP) samples (n = 26), where ctDNA was expected only in non-cured patients, PyDREAMS detected ctDNA in 5 patients, including 3/9 with later recurrence, while Shearwater detected none. Together, these results show that PyDREAMS improves evaluation of ultra-low-frequency tumor signals through unified read-level modelling of SNV and indel background error.

bioinformatics↗

bioRxiv2026

Combining mutation detection with fragmentomics features leads to improved tumor-informed ctDNA detection

Liquid biopsy through circulating tumor DNA (ctDNA) analysis enables non-invasive detection of minimal residual disease (MRD) and early identification of cancer relapse, facilitating timely clinical intervention. However, detecting ctDNA in plasma samples with low tumor burden remains challenging due to the scarcity of mutant molecules, the background noise of sequencing errors and somatic mutations in normal cell-free DNA (cfDNA). Here, we present a mutation-informed fragmentomic framework and evaluate it on 90 stage III colorectal cancer patients with three years of follow-up. Using 712 serial whole-genome sequenced cfDNA samples (30x) with matched whole-genome sequencing of tumor tissue and germline DNA from buffycoat for each patient, we collected cfDNA fragments spanning tumor-derived somatic mutation positions and compared fragmentomic characteristics of mutation-bearing and non-mutated cfDNA fragments within the same sample. By leveraging fragment length and fragment end-motif patterns, our approach can distinguish cancer-positive from cancer-negative plasma samples without requiring model training or panel-of-normals calibration. The method achieved AUCs of 0.863 and 0.74 using fragment length and end motif features, respectively, and 0.871 when combined, outperforming tumor fraction estimates based on the frequency of mutated fragments (AUC=0.832). Integrating fragmentomic features with tumor fraction further improved performance, yielding an AUC of 0.873, indicating complementary signals between fragmentomic patterns and mutation burden. Aggregated analyses revealed ctDNA-specific patterns, including fragment shortening, motif enrichment of A/T ends, and depletion of C/G ends, directly linking fragmentomic features to tumor-derived cfDNA. Overall, mutation-informed fragmentomic profiling improves ctDNA detection beyond counting mutant reads and provides a scalable, training-free strategy for MRD assessment and early relapse detection while offering mechanistic insights into tumor-specific cfDNA biology.

bioinformatics↗

bioRxiv2025

A reference-free strategy for circulating tumor DNA detection from whole-genome sequencing data

Circulating tumor DNA (ctDNA) is emerging as a promising biomarker for postoperative monitoring of cancer patients. Precise estimation of circulating tumor fraction is crucial for evaluating treatment effects and timely detection of disease recurrence. All current ctDNA detection methods that utilize whole-genome sequencing (WGS) data rely on the reference genome alignment of sequencing reads and often apply separate tools for detecting different variant types. However, various bioinformatic analysis confounders and the application of external variant calling tools could be avoided by analyzing k-mers from unaligned sequencing reads. While k-mer-based methods have successfully been applied for somatic variant validation and detection, the potential of k-mer-based ctDNA detection is unexplored. We have developed a tumor-informed reference-free ctDNA detection tool called ctDNAmer that detects tumor-specific somatic variation directly from unaligned sequencing data by identifying k-mers unique to the tumor DNA. ctDNAmer detects variant information across the genome by comparing the primary tumor and germline WGS data and accounts for sample-specific germline variability and technical noise in the same framework. We tested the utility of ctDNAmer for tumor fraction estimation on postoperative plasma cfDNA WGS data (mean sequencing depth ~28x) from 90 stage III colorectal cancer patients with three years of follow-up. The tumor fraction (TF) estimates agreed with the available clinical information and ctDNA was detected in 77% (17/22) of recurring patients with a median lead time of 8 months compared to radiological imaging. We further validated ctDNAmers tumor fraction estimates based on a comparison with the mean cfDNA allele frequencies of somatic clonal SNVs identified from aligned primary tumor sequencing data. The TF estimates showed a strong Pearson correlation of 0.897 with the mean allele frequencies and improved ctDNA detection results across samples with an AUC of 0.79 compared to 0.75 if the mean allele frequency of clonal mutations is used.

bioinformatics↗

bioRxiv2024

Accurate calling of low-frequency somatic mutations by sample-specific modeling of error rates

Calling rare somatic variants from NGS data is more challenging than calling inherited variants, especially if the somatic variant is only present in a small fraction of the cells in the sequenced biopsy. In this case, having a good estimate of the error rate of a specific base in a particular read becomes essential. In paired-end sequencing, where some DNA fragments are shorter than twice the read length, the overlapping regions of the read pairs are an ideal resource for training models to discern context-dependent base error rates, as any discordant bases in the overlaps must be caused by a sequencing error or an alignment error. We have created a new tool named BBQ (an acronym for Better Base Quality) that uses overlapping reads to estimate the error rate conditional on the mutation type, sequence context, and base quality. We also estimate how much the error rate of concordant bases in overlapping reads is decreased compared to bases in non-overlapping reads. Results show that overlapping reads can remove sequencing errors induced by DNA damage and that the increased quality of overlapping reads differs between samples and mutation types, reflecting different damage patterns between samples. We use the error models to call rare somatic variants. Sequencing data from a testis biopsy and a cell-free DNA sample serve as a proof-of-concept for rare germ cell mutation calling and for detecting rare cancer mutations. We find that using the sample-specific error models of BBQ allows us to call rare somatic variants with fewer false positives than existing tools such as Mutect2 and Strelka2.

bioinformatics↗

bioRxiv2023

Error-corrected deep targeted sequencing of circulating cell-free DNA from colorectal cancer patients for sensitive detection of circulating tumor DNA

IntroductionCirculating tumor DNA (ctDNA) is a promising biomarker, reflecting the presence of tumor cells. Sequencing-based detection of ctDNA at low tumor fractions is challenging due to the crude error rate of sequencing. To mitigate this challenge, we developed UMIseq, a fixed-panel deep-targeted sequencing approach, which is universally applicable to all colorectal cancer (CRC) patients. MethodsUMIseq features UMI-mediated error correction, exclusion of mutations related to clonal hematopoiesis, a panel of normals for error modeling, and signal integration from single-nucleotide variations, insertions, deletions, and phased mutations. UMIseq was trained and independently validated on pre-operative (pre-OP) plasma from CRC patients (n=364) and healthy individuals (n=61). ResultsUMIseq displayed an area under the curve surpassing 0.95 for allele frequencies (AF) down to 0.05%. In the training cohort, the pre-OP detection rate reached 80% at 95% specificity, while in the validation cohort, it was 70%. UMIseq enabled the detection of AFs down to 0.004%. To assess the potential for detection of residual disease, 26 post-operative plasma samples from stage III CRC patients were analyzed. Detection of ctDNA was associated with recurrence (p =0.08). ConclusionUMIseq demonstrated robust performance with high sensitivity and specificity, enabling the detection of ctDNA at low allele frequencies.

cancer biology↗

bioRxiv2022

DREAMS: Deep Read-level Error Model for Sequencing data applied to low-frequency variant calling and circulating tumor DNA detection

Circulating tumor DNA detection using Next-Generation Sequencing (NGS) data of plasma DNA is promising for cancer identification and characterization. However, the tumor signal in the blood is often low and difficult to distinguish from errors. We present DREAMS (Deep Read-level Modelling of Sequencing-errors) for estimating error rates of individual read positions. Using DREAMS, we developed statistical methods for variant calling (DREAMS-vc) and cancer detection (DREAMS-cc). For evaluation, we generated deep targeted NGS data of matching tumor and plasma DNA from 85 colorectal cancer patients. The DREAMS approach performed better than state-of-the-art methods for variant calling and cancer detection.

bioinformatics↗

bioRxiv2022

Machine learning guided signal enrichment for ultrasensitive plasma tumor burden monitoring

In solid tumor oncology, circulating tumor DNA (ctDNA) is poised to transform care through accurate assessment of minimal residual disease (MRD) and therapeutic response monitoring. To overcome the sparsity of ctDNA fragments in low tumor fraction (TF) settings and increase MRD sensitivity, we previously leveraged genome-wide mutational integration through plasma whole genome sequencing (WGS). We now introduce MRD-EDGE, a composite machine learning-guided WGS ctDNA single nucleotide variant (SNV) and copy number variant (CNV) detection platform designed to increase signal enrichment. MRD-EDGE uses deep learning and a ctDNA-specific feature space to increase SNV signal to noise enrichment in WGS by 300X compared to our previous noise suppression platform MRDetect. MRD-EDGE also reduces the degree of aneuploidy needed for ultrasensitive CNV detection through WGS from 1Gb to 200Mb, thereby expanding its applicability to a wider range of solid tumors. We harness the improved performance to track changes in tumor burden in response to neoadjuvant immunotherapy in non-small cell lung cancer and demonstrate ctDNA shedding in precancerous colorectal adenomas. Finally, the radical signal to noise enrichment in MRD-EDGE enables de novo mutation calling in melanoma without matched tumor, yielding clinically informative TF monitoring for patients on immune checkpoint inhibition.

genomics↗

bioRxiv2019

First analysis of behavioural responses of humpback whales (Megaptera novaeangliae) to two acoustic alarms in a northern feeding ground off Iceland

Mitigating cetacean entanglement in fishing industries is of global interest. Strategies include the use of acoustic alarms to warn whales of fishing gear. For baleen whales, responses to acoustic alarms are poorly understood. This behavioural response study compared the behaviour of humpback whales (Megaptera novaeangliae) in their feeding grounds off Iceland prior to, during, and after exposure to a low-frequency whale pinger (Future Oceans) and a high-frequency seal scarer (Lofitech ltd.). Linear mixed effects models and binary generalized linear mixed effects models were used to analyze the effect of the alarms on surface feeding, swimming speed, breathing rate, directness and dive time. We observed a significant decrease in surface feeding and a significant increase in swimming speed during exposure to the whale pinger. Changes in dive time between the phases of a trial differed significantly between individuals indicating that responses may depend on individual or behavioural state. We did not find any significant reactions in response to the seal scarer. In addition to the experimental exposures, a trial of whale pingers on a capelin purse seine net was conducted. Results from this trial showed that whales entered the net from the bottom while the pingers were attached at the top, but the encircled whales were able to locate an opening free of pingers and escape without damaging the net. Our results suggest that whale pingers may be a useful entanglement mitigation tool in humpback whale feeding grounds given that a reduction in feeding around nets likely reduces the risk of whales swimming through them. Pingers may also minimize net damage if whales are encircled by aiding the whales in finding their way out. However, given the uncertain long-term consequences of the behavioural changes reported here, whale pingers are most advisable for short-term use in conjunction with other entanglement mitigation measures.

animal behavior and cognition↗

bioRxiv2018

Abundance of whales in West and East Greenland in summer 2015

An aerial line transect survey of whales in West and East Greenland was conducted in August-September 2015. The survey covered the area between the coast of West Greenland and offshore (up to 100 km) to the shelf break. In East Greenland, the survey lines covered the area from the coast up to 50 km offshore crossing the shelf break. A total of 423 sightings of 12 cetacean species were obtained and abundance estimates were developed for common minke whale, from now on called minke whale, (Balaenoptera acutorostrata) (32 sightings), fin whale (Balaenoptera physalus) (129 sightings), humpback whale (Megaptera novaeangliae) (84 sightings), harbour porpoise (Phocoena phocoena) (55 sightings), long-finned pilot whale, from now on called pilot whale, (Globicephala melas) (42 sightings) and white-beaked dolphins (Lagenorhynchus albirostri) (50 sightings). The developed at-surface abundance estimates were corrected for both perception bias and availability bias if possible. Data on surface corrections for minke whales and harbour porpoises were collected from whales instrumented with satellite-linked time-depth-recorders. Options for estimation methods are presented and the preferred estimates are: Minke whales: 5,095 (95% CI: 2,171-11,961) in West Greenland and 2,762 (95% CI: 1,160-6,574) in East Greenland, fin whales: 2,215 (95% CI: 1,017-4,823) in West Greenland and 6,440 (95% CI: 3,901-10,632) in East Greenland, humpback whales: 993 (95% CI: 434-2,272) in West Greenland and 4,223 (95%CI: 1,845-9,666) in East Greenland, harbour porpoise: 83,321 (95% CI: 43,377-160,047) in West Greenland and 1,642 (95% CI: 319-8,464) in East Greenland, pilot whales: 9,190 (95% CI: 3,635-23,234) in West Greenland and 258 (95% CI: 50-1,354) in East Greenland, white-beaked dolphins 15,261 (95% CI: 7,048-33,046) in West Greenland and 11,889 (95% CI: 4,710-30,008) in East Greenland. The abundance of cetaceans in coastal areas of East Greenland has not been estimated before, but the limited historical information from the area indicate that the achieved abundance estimates were remarkably high. When comparing the abundance estimates from 2015 in West Greenland with a similar survey conducted in 2007 there is a clear trend towards lower densities in 2015 for the three baleen whale species and white-beaked dolphins. Harbour porpoises and pilot whales however, did not show a similar decline. The decline in baleen whale and white-beaked dolphin abundance is likely due to emigration to the East Greenland shelf areas where recent climate driven changes in pelagic productivity may have accelerated favourable conditions for these species.

ecology↗